TL;DR: Choose where your files live based on who or what needs access, the devices you use, and whether you can move the work later. Check that access and test an export before moving everything in.

My organized briefcase still needed me

I've always been organized. Papers, years of tax returns, old statements, bank statements, and canceled checks. I kept it all.

So when Suze Orman came out with this organization system, a little briefcase with divider folders inside for all of the different documents you needed to keep for good financial health, I bought it. I was all in, and it was great.

I had this plastic briefcase, and I was able to put a lot of stuff in it and store it there.

The only problem was access. If anyone needed something, I had to physically go in, pull out the paper, send the paper to where it needed to be, and then put the paper back.

When it was just me, it was fine. But when it was my husband or anyone else, they had to learn my system. And my system was this plastic briefcase with these nice little neat divider sections.

There were so many papers in there that trying to find what you needed, even in the sections, became a futile effort. It took forever. Well, not forever, but it took a good while to sort through all of the papers.

Because even though it was sectioned, there was still a lot in there. There were papers for our home, and then papers for work, and then papers for the kids, and then papers for the family and other stuff. It was a lot of papers.

Then scanning became a thing. You could scan your documents and put them on a hard drive. That was the new thing, right? So I bought a portable scanner and started to scan everything. And then they went onto CDs, because back in the day that was the storage of choice. You burned stuff onto CDs.

So now, instead of having to root through papers, I had to root through CDs.

Then we got into memory sticks and hard drives, and that gave me even more options. Instead of CDs that could get scratched or broken, I had a hard drive or a memory stick that was easier to share without damage.

Fast forward all these years, and we have cloud storage and digital copies. Keeping information has never been easier, but there's still some separation.

Now my financial records are automatically pulled into my bookkeeping software, for both personal and business. It also brings in my credit cards and my investments, so I have everything in one place. It just funnels in electronically, instead of me having to open up each application and try to find the statement or the information.

With services like Zapier and Make, MCPs, and even API connections, it's easier to get information from point A to point B, if you know what that route is.

And that was the same problem I started to encounter when I started to work with multiple AIs and what got stored where.

Two AIs needed the same work

Content Lab is the process I use to create these articles. Claude and Chat both help me create and refine them.

At first, it was Chat for this and Claude for that. A lot of it came down to allotments: I have this many tokens in this AI and this many tokens in that one.

The biggest change came when I started to really use both AIs as part of the process. I wanted to take the strengths of both of them, or even other sources like Midjourney and Gemini and the other tools I had access to.

So I started to think in who writes better, who thinks better, who researches better.

Maybe you're there too. One AI for writing, another for research, and your work spread out between them.

When working with two AIs separately, there was that separation. One picked up where the other left off, as a clear break.

But now there is back and forth: one for assembling the context, one for research and development, then the brief, move to another for writing, and back for images.

It's the back and forth to determine who has what role, and is everyone on the same page, instead of each hoarding and doing their own thing.

It's like working with a team who each have a home office.

And now AI is moving a lot of stuff to its own cloud. ChatGPT's Spaces let me share work with a person. But that doesn't tell me whether Claude can work directly with the same information there. And on Claude's Pro and Max plans, new Cowork sessions and scheduled tasks now run in the cloud, with the work saved to your Claude account. Older tasks that started on your computer stay there.

So I need to check two things: can another AI use the same work where it lives, and can I download or export a copy to use somewhere else? Getting a copy out is different from both AIs working with the same current information. That's the back and forth I built my processes around.

So I don't want to do all kind of work in one area and then be stuck there.

✳︎ ✳︎ ✳︎

What to think about: Who or what needs your information?

When I started to pick AIs for their strengths, I realized that the whole process has to be agnostic as far as the documentation and the information.

It isn't just about the storage where the files are held, but also how and who can access them.

And who needs access can change from one piece of work to the next:

• Sometimes it's just yourself.

• Sometimes it's yourself and AI.

• Other times it's outside sources. Your team, a colleague, or even a subcontractor. Someone who is going to help you with this information.

It has to be available to whoever needs it. For Content Lab, it was just two AIs. I didn't need an outside source to have access to it.

Where the files live changes who can reach them. Cloud storage like Google Drive is kind of your protected space, but you can share it. Say I'm working on an article, I'm citing a source, and I want that source to review their section. I can send that document to them where they can read and access it.

That's different than a cloud storage that's attached to AI, or that is part of AI. If something requires more than one AI, like how the Content Lab works, I need to check whether the other AI can reach that information or whether I have to move a copy between them.

And some information you may not want in the cloud at all. On your own computer, you can choose which files to give AI access to. But keeping a file there doesn't mean its contents stay off AI servers when the AI reads it. Where you store the file and where the AI processes its contents are two different things.

So when I figure out what information to store where, I immediately think: who or what is going to need access to this? And does it lock me into a particular tool?

✳︎ ✳︎ ✳︎

What can go wrong: When the tool changes and your work can't leave

I've seen this with a lot of tools. You start out on a free tier, or maybe a low-cost paid tier, and then the price increases.

The problem becomes, if you're locked in, you either have to pay that price or lose the information.

Now sometimes it's just nominal or minimal changes, which you have to expect, especially as costs increase. But there are times where I've seen programs where the increase was a lot more.

Adobe was a big thing with that. They get you in. I got in the door at a promotional price, and at the end of that term, the price jumped.

Now I know you have to factor in the benefits that you get. But the problem is, you're brought in at a promotional rate, and then it jumps. So now you have this extra added expense every month for something that you've gotten used to and comfortable with.

In the case of Photoshop, there wasn't a whole lot of options for photo editing, so you kind of got locked into it. If you wanted that functionality, you had to pay that price.

The same thing can happen when your work lives in one AI's cloud. If something changes and you can't move your work out, you're back to paying the price or losing the information.

✳︎ ✳︎ ✳︎

What comes next: Check before you move your work in

So before you move your work into an AI space or any tool space, check these three things:

• Who or what needs access to it? It might be just you, you and AI, or outside sources.

• Can you reach it from all of your devices—and can the AI reach it when you work from those devices?

• Can you take it with you if something changes?

First, ask. Paste this into the AI before you save your work there:

Before I save my work here, tell me: Where is it stored? Can I download or export it, and in what format? Can another AI or another person use it? Can I reach it from my phone and tablet? If I use AI on those devices, can it read and update the information too?

An AI isn't always right about its own features, so check its answer against the tool's help pages.

Then, test it. Put one small piece of work in, and then try to get it out. Download or export it, and open it in a different tool. From the devices you use, also check whether the AI can read and update that piece of work.

If you can't get it out, or the AI can't reach it where you need to work, you know before you move everything in.

✳︎ ✳︎ ✳︎

Where I chose to keep the work

The reason I'm keeping my work in Obsidian, a free app that reads and organizes Markdown files, is because I use more than one AI. I didn't want it locked into one space where the others didn't have access to what was going on.

I started with my hard drive, because I wanted to keep it accessible to the AIs, where I could share it easily, be it an uploaded file or whatnot. But a lot of the information I was saving needed two things: Markdown, and a way to tag and sort information.

Obsidian could read the Markdown. I could use the fields to tag and categorize things. And it's easy for AI to understand.

But it isn't so much about Obsidian. It's Obsidian with iCloud. My Obsidian folder is set up in iCloud, so it's accessible from my devices, because all of my devices have a connection to iCloud. And if I need to share a file, I can easily do that.

Realistically, it could be Obsidian with Drive, or Obsidian with Dropbox, or Obsidian with my hard drive.

Basically, it's just a file system and reader:

• I have my choice of where I want to store the information.

• I have a choice over who and what I share that information with.

• I have a choice of how I want to structure that information.

It's not completely dictated for me.

I'm sure that I have files in the AI cloud. But by storing your work in an area that is specific to your needs, and that is portable, you avoid getting stuck.

Before you move your next piece of work into an AI's space, run the prompt and the test. Then tell me what you found. Could you get your work back out?

You don’t have to move everything at once. Start with one piece of work and see whether the setup fits how you actually use it.

✳︎ ✳︎ ✳︎

Help with your next decision

Where your work lives is one decision. If you want help working through the next one, the Decision Toolkit pairs a workbook with companion AI skills.

Get access to the Decision Toolkit, available in the perks.

TL;DR: I moved DigiNav from WordPress to Astro because a different foundation fit what I wanted the site to do and how I work. Before you change your own site's platform, look at its functions, the visitor's experience, and what you can update and maintain. Staying where you are may be the right decision for now.

What would change about the work?

I can still remember clear as day when I had the opportunity to work remotely. This was before 2000 when the internet connections still meant dial-up. The idea was intriguing and I was curious about all the ways that it would work. But the question started almost immediately as soon as I pictured me working from a home office.

Would I be able to keep up the work and hit my numbers? How would it be different without the office politics and giving people the feedback that they needed? Did I really want to work from my basement every day?

I weighed the pros and cons and realized that some of my concerns weren't really concerns at all. My husband worked for the phone company, so adding the dedicated line was possible. And I had the tech support close by. Plus I had room to create a secure office. The part I couldn't check off so easily was the work itself.

I had a flow. I would have to learn a completely new way to work and unlearn parts of the process that now came so easily.

But me being me, I decided to give it a try. We set up the office and added the phone line. Now it was time to connect online and get some work done. And I have to say at first it was painfully clunky. The system was slow, the server was even slower. We found that evenings worked better when the regular staff wasn't connected.

I also had to change how I entered information because of the way that the data moved and was saved. In the end, I made it work. And it actually was an enjoyable experience. And then we added two more people to join the remote work program. And even though the answers came not from a list of pros and cons, but actually how things were done, I had to see how the change would affect the work and then adjust when the first plan didn't meet the reality of my expectations.

✳︎ ✳︎ ✳︎

Hi, I’m Lee. Welcome to DigiNav Compass Signal. Around here, I share what happens when AI and technology meet a real business—the useful parts, the messy parts, and what you can take into your own work.

Subscribe now

Today, I’ll be showing you how I decided whether DigiNav needed a different foundation, and the questions you can use to examine your own site before you make a plan.

✳︎ ✳︎ ✳︎

I could make WordPress work

This decision came back to me when I started looking at moving DigiNav from WordPress to a code-based site. Only this time, the pros and cons were completely different. When I first built websites, I used HTML, CSS, and other code. Whether I could hand-code a site wasn't the question. I started using WordPress because it was faster and easier to build. Over the years, I became really good at it. I built my processes around it, and I still run the WordPress agency today.

That was the first big factor in the decision. How would I be perceived as a WordPress agency if I moved DigiNav away from WordPress? I also wanted to ditch the database and make DigiNav more interactive and fluid. WordPress could do those things, but getting there wasn't necessarily easy. If I was going to custom-code the options anyway, why add another layer in between?

That's where the self-doubt came in. Some of it was removing the cape from the work I was doing with WordPress and putting on the cape of using code and AI to make the website. I also wondered whether I was making the move more complicated in my own mind because I was so familiar with WordPress and the processes I had built around it. I wasn't just choosing a new system. I would have to change the way I worked with it.

✳︎ ✳︎ ✳︎

Name the job before you ask about platforms

The first question was whether moving DigiNav off WordPress was a better option at all. WordPress could do much of what I wanted. I needed to look at what a move would make possible, what it would require, and whether staying would serve the work just as well.

For DigiNav, the useful starting point was specific: articles needed to work with the Waypoints, the visitor's path needed more interaction, and the content needed to be easy to find. Commerce could stay in ThriveCart. A simple form could handle basic collection. A content database was something to leave behind, not another system to maintain.

If you are weighing a move, try naming four things before you compare platforms:

• What must keep working on the current site?

• What do you want to make easier or add next?

• What can stay in a service outside the site?

• What will you have the time, skill, and support to update and maintain?

You can use an AI chat to help examine those answers. Here is a prompt to adapt, not one to follow word for word:

I am thinking about moving my website from [current platform]. First, ask me what the site must do now, what I may want it to do next, and what I do not want to maintain. Then compare staying where I am with up to two other ways to build it. For each option, explain what would have to move, how I would update the site, what I would maintain, and what could become harder. Use plain language. Do not choose a platform or make a build plan yet.

Looking at the whole job—what the move would make possible, how I would move forward, and how I would update and maintain the site—helped me narrow the options to three. I compared their pros and cons from there, and Astro was the choice that made sense for DigiNav.

My coding and technical background helped me make sense of those choices. If you don't have that background, include it in what you ask AI. You can add a line like: “I don't code, and I'm not very technically savvy. Factor that into this decision too, including what I would need to learn or get help with to update and maintain the site.”

✳︎ ✳︎ ✳︎

The platform is one decision

A platform comparison shows what a new site can do. It cannot tell you what people will experience when they use it. If you are reworking a website, you need to check more than the framework:

• Usability: Can someone find what they came for?

• User interface (UI): Is it clear what they can do on each page?

• User experience (UX): Does one step lead naturally to the next?

This was one of the biggest focal points for the DigiNav migration. I could have taken the WordPress site, converted it to Astro, and called it a day. But I didn't feel like users were getting as much out of the experience as they could. Once the framework question had an answer, it was time to address the user part of the work.

✳︎ ✳︎ ✳︎

Step one is to name the decisions

Before you make a migration plan, get clear on the decisions the plan will have to carry. For DigiNav, choosing a possible framework was only one part. The site also needed to work better for the people using it.

Here are the questions to address before you commit to a move:

• Timing: Do you have room to give this decision the time it needs?

• Platform: What does the current system make easy, and where does it add work? Would you stay with it or move to a different one? What would a new platform need to support?

• Size of the move: Is this a brochure site—a small site that mainly tells people what you do and how to contact you? That usually leaves fewer decisions to make. A site with a store, membership, or custom functions has more to account for before you choose how to move it.

• Functionality: What must keep working? What new functions will the site need, and what can live in a service outside it?

• Visitor experience: Where do people get stuck? What should be easier to find, understand, or do? If you want more interaction, what job will that interaction serve?

• Accessibility: How will people use the site with a keyboard, a screen reader, or on a small screen? How will you check that the move protects access?

• Care after launch: Who will update the content, maintain the site, and fix what breaks? What time, knowledge, and help will that take?

✳︎ ✳︎ ✳︎

Staying where you are may be the better choice for now if you cannot give the decision enough time, you are in the middle of a business pivot, or your site has functions that need more thought before you move them. You may also need time to learn the tools you would use to build and maintain a new site. Just check whether that is a real limit or whether fear of the unfamiliar is making the decision for you.

By the end of this first phase, I had made the decision to move DigiNav off WordPress and build it with Astro. I still get that queasy feeling that I’m abandoning WordPress, even though I still run a WordPress agency. Both things can be true: WordPress still serves my clients, and DigiNav needs a different foundation for what I want to build.

My coding background helped me judge the options. I also knew I could use AI to help with much of the build, which would give me more room to focus on how DigiNav grows and how people use the site. The first small changes have been easy to make. How the site holds up as it grows is something I’ll keep checking.

That was the outcome of phase one: a decision with a clear reason behind it. In the next article, we’ll look at how the site’s needs led to the technical choices, including Astro.

But some of my favorite uses take a few minutes and are generally random.

This week, I asked about a sweater, got help with a word game, and tried to understand why an image tool kept adding people I never asked for.

Here are five questions I asked AI this week and what came out of them.

✳︎ ✳︎ ✳︎

1. Sometimes I want another opinion.

I had a casual sundress, a sweater, and shoes picked out for a wedding. I couldn’t decide whether the sweater worked or made the whole thing look too formal.

Instead of texting someone a photo and waiting for an answer, I asked AI.

The prompt that started it:

“Does this sweater look OK with this sundress?” [insert two pictures]

My life lesson:

This was a small decision. A second opinion was enough to help me make it.

✳︎ ✳︎ ✳︎

2. AI is perfectly acceptable for cheating at word games.

At one point this week, my biggest problem was staring at a pile of letters and failing to see the word hiding in them.

So I started feeding the letters to AI.

STATUS. MOONLIGHT. CLASSROOM. ARTICLE.

I won’t pretend this was productive. But it was useful at that exact moment.

The prompt that started it:

“Use these letters to spell a word, don't add or subtract any: asstut”

My life lesson:
Next time you play Scrabble, sneak off and get a second opinion.

✳︎ ✳︎ ✳︎

3. Two weird failed orders can give you something to investigate.

A client had two failed orders come through her online store with strange names and even stranger notes.

My first thought was: Is someone testing stolen credit cards?

The prompt that started it:

“I just had two failed orders come through my cart… was it testing credit cards?”

That question helped me investigate card testing, how bots use checkout forms, and what Stripe was already blocking. It also led me to check where Cloudflare Turnstile could add another layer of protection.

A random “what the hell is this?” became an actual fix.

My life lesson:
Ask about what you suspect. Then check whether the evidence supports it.

✳︎ ✳︎ ✳︎

4. Sometimes the problem with the picture isn’t your prompt.

I spent far too much time trying to get Midjourney to create a homepage image without putting people in it.

More specifically: men.

Tell it “no people.” There’s a man.

Try a different scene. Man.

Change the concept. Hello again, random man.

Eventually, I stopped asking for another image and asked about the problem itself.

The prompt that started it:

“Why does Midjourney put men in every f*cking picture, even if you tell it no people?”

That led to a conversation about patterns in training data and why an image generator might keep adding something I didn’t want.

It also gave me a different approach to try: Describe a scene where a person has no reason to be.

My life lesson:
Sometimes a question and a cuss word help you stop repeating the same failed approach.

✳︎ ✳︎ ✳︎

5. “Can AI see this?” is becoming a more useful question.

I wanted to connect AI to Cloudways so I could ask questions about my servers. I was tired of digging through dashboards and trying to interpret everything myself.

The prompt that started it:

“Do you have an MCP for Cloudways servers?”

That sent me into connection tools, access tokens, and permissions. Knowing about Cloudways is one thing. Having permission to inspect my account is another.

It also gave me a better question to ask: What information can I safely give AI access to so it can help me understand what’s happening?

My life lesson:
Before spending time collecting information by hand, ask whether AI can access the source and what permission it needs.

✳︎ ✳︎ ✳︎

None of these questions needed a sophisticated prompt. I asked the question that was in my head, then followed up when I needed more.

That’s an easy way to get more comfortable using AI. Ask about the small, weird, ordinary things that catch your attention.

Sometimes you get an answer. Sometimes you find a different question worth asking.

And sometimes you finally figure out what word you can make from asstut.

TL;DR: A working process can still have steps you no longer need. When your tools or needs change, review those steps, keep what works, and test one small improvement. The prompt below helps you work through that review.

The workaround was creating more work

For several months I've been working on an email triage system so that things coming into my inbox were not getting lost in the shuffle between the agency that I still run and here at DigiNav.

Originally, the way that I set up the triage was that the sorting would be handled by Claude to decipher what needed my attention, what needed my response, what was just noise, and so forth.

The problem with that was that in order to get the emails to a place where they could be sorted, I had to forward my agency emails, which were in another Google Workspace account, into my main email account that was hooked up with the AIs.

And that caused a small, tiny friction: two copies of the same message in different email accounts.

✳︎ ✳︎ ✳︎

✳︎ ✳︎ ✳︎

The short version is that I did three things:

1. I noticed a change in a tool I already use.

2. I matched it to a small friction I had been living with.

3. After the fix, I looked at the next step to see if that could improve too.

If you have a process that works but still has a little extra handling in it, this is for you.

When the help account was reviewed, the copy was still there, and it made it look like it was either sitting in the inbox or unfinished.

Still, the sorting system was useful. It was working. Emails were getting sorted, nothing was getting lost anymore, and I just went and cleaned up the secondary email once a week.

But it was still a small friction.

Then I was watching a What's Happening in AI YouTube video, and I noticed there was a change in the way the connections were now working in ChatGPT that gave me a way to fix this system. I follow a handful of creators on YouTube who only cover changes that matter for a business like mine. Their recaps help me funnel through all the changes without having to read a bunch of documentation.

✳︎ ✳︎ ✳︎

One change made another possible

The one change was that now, in ChatGPT, you can connect multiple Google Workspace accounts. And so that solved tiny friction number one.

Instead of forwarding emails to one account and having duplicate copies, I could now work from the correct email accounts. And this led to the change in where the sorting happened. That meant I could move the scheduled sorter from Claude and stop relying on the forwarded copies to get everything into the process.

So first things first, the sorting rules that came from Claude covered the basics of what needed attention, what could be kept, and what was noise. That had already been worked out.

The change was where the sorting would reach the email. Each email should stay in its appropriate account when it arrived and then be sorted.

A few tiny tweaks to the sorting process, and that was an easy change.

And the best part was I didn't have to take a lot of time to figure it out. I put the old information or the instructions into ChatGPT and I said, let's fix this, and it did it all.

✳︎ ✳︎ ✳︎

The morning brief could run automatically too

So now that I had the proper way to sort two email accounts, I noticed there was another step I was doing by hand that I could automate to save time.

And that was the morning briefing. I know there's a lot out there about how easy it is to create a morning briefing, but my problem was the sorting was done in one AI and the briefing was done in another.

So I did the same thing as when I brought the sorting over: I asked ChatGPT to review the email briefing process and recommend the best way to automate this without breaking the next step in the system, which is my daily planning.

So by bringing the sorting over, it also gave me a way to schedule the brief too. Now I have a weekday morning task set up to do that review, and it's ready for me when I am ready to start my day. Now, don't get me wrong, it's not a done deal. I'm still refining the system, but this removed another step that depended on me remembering to start it.

✳︎ ✳︎ ✳︎

What has changed since you built your system?

This process isn't any different than any other process in your business. You build it, use it, and find ways to improve it. That applies to how you bring on clients, prepare for meetings, or even sort your emails.

Sometimes it's the tools that change, sometimes your needs change, and sometimes you notice that a step that you keep doing no longer serves the purpose it once did. And that is especially true as new features and benefits are added to AI and automation technology.

The forwarding workaround and the weekly cleanup were a small friction I had accepted because the sorting system was doing its job.

But when the connection option changed, I had a reason to not only improve the beginning of the sort system, but I had a reason to review other steps after that.

And you can do the same with a process you already use. Pick one that works but still requires a little bit of extra handling or has a small friction in it.

You don't need to have the solution figured out before you review it.

Use this prompt to review one process

Copy this into the AI you already use:

Help me review one process in my business and find one useful improvement.

Ask me one question at a time. Start by asking which process I want to review and what it needs to accomplish. Then help me walk through how it works now.

Find out which parts work well, where I repeat work or make manual corrections, and what has changed since I set it up. Consider changes in my needs, workload, people, tools, or available features.

Use my answers to suggest the smallest useful change. Explain what it would remove or improve, which working parts should stay, and what later steps depend on it.

After that, look at the steps that come next. Ask whether the change makes any other manual step unnecessary, and name it if it does.

Give me a simple way to test the change and tell whether it helped. If your suggestion depends on a tool feature, verify that feature or clearly tell me what I need to check.

Keep the review focused on this one process. Do not assume I need a new tool or a full rebuild.

As you make those changes, keep the context that helped you make them. What was working? Why did you change this part? What still needs to be tested? That gives you and your AI a place to start when you come back to refine the process again.

✳︎ ✳︎ ✳︎

If you want help keeping that information together, my Pick Up Where I Left Off subscriber perk gives you three AI prompts and simple project templates to record where things stand, the background behind your decisions, and what remains to be done. You can use it alongside the review prompt above to keep the context as your system changes.

The package is available to paid subscribers. Use it to leave yourself a clear place to return, so the next improvement can build on what you've already worked out.

TL;DR: A respected expert can still be the wrong AI advisor for your business. Start with the job you need help with, then check the person's documented work, business fit, blind spots, and whether the perspective adds something useful before giving it a seat.

I chose the people before I named the jobs

Last week, I shared why I built an AI Board of Advisors and why it needed boundaries. I work solo, and I missed the mentors and mastermind groups I once turned to when I got stuck inside my own thinking. The Board was my way of bringing those outside perspectives back into the room.

Hi, I’m Lee. Welcome to DigiNav Compass Signal. Around here, I share what happens when AI and technology meet a real business—the useful parts, the messy parts, and what you can take into your own work.

Subscribe now

Today, I’ll be showing you what I got wrong when I chose my first advisors and the five questions I now use to decide whether an AI advisor fits the business it is supposed to help.

The first version of my AI Board of Advisors started with people whose work I respected. Some I knew and learned from over the years. Others were before my time, but they were just as strong and knowledgeable.

But when I put the panel in front of real business decisions, the gaps became obvious. No one covered my biggest weakness: sales. And I quickly found out that the growth and monetization advice leaned on high-volume creator models and bigger teams that did not fit the small, relationship-led business I was building.

The problem was not that these people had nothing useful to teach me. The problem was that I had started with the people. I needed to start with the jobs.

Someone can be a respected expert and still be the wrong advisor for the work in front of you. Their method can be sound and still depend on a business model, pace, team, audience, or definition of growth that does not fit yours. They may belong in your research without earning a seat at your table.

That changed the question from, “Whose ideas do I admire?” to, “What job needs to be done, and whose documented method fits a business shaped like mine?”

Five questions before an advisor earns a seat

When I rebuilt the board, I stopped treating advisor selection as a search for the smartest person in the room. I started testing each possible advisor against the job I needed done and the business the advice had to serve.

You can use the same test with one person. You do not need to build a full board first.

1. Name the job. Start with one recurring decision, tension, or part of the business where another perspective could help. Be specific. Positioning, sales, audience growth, and monetization are connected, but they are not the same job. I could not fix a sales gap by adding another general business thinker.

2. Check the evidence. Look for documented work that directly covers that job: books, articles, talks, interviews, or other public sources that show how the person thinks and works. A large following, a few good quotes, or an AI summary does not prove that the method fits the role. That documented work becomes the source for the advisor profile the AI uses.

3. Check the business fit. Good advice can still be wrong for your business. Look at what the method assumes about team size, time, money, audience, pace, delivery, and growth. If the advice depends on publishing every day or handing the work to a bigger team, it may not fit a business built around limited capacity and trusted relationships.

4. Record the blind spots. Every method has limits. Ask where it could stretch beyond its real expertise, import a business model you do not want, or create work you cannot support. Those limits belong in the advisor profile too. They tell the AI when that perspective should stay out of the decision.

5. Make the advisor earn the seat. The person should add something you do not already have. More agreement is not always more useful. Sometimes you need a different view. Sometimes you need better evidence. And sometimes the honest answer is that this person belongs in your research, but not on your advisory board.

At the end of the tests, the result is not based on a simple yes or no. The advisor may be okay for a small test or remain as a research source. They could need more evidence, or they could just not fit at all. Leaving an empty seat is better than filling it with someone whose advice was never built for the business you're building.

Start with one advisor, not a full board

The first thought is to build a complete and full board. But before you do that, pick one part of the business where another perspective would be the most helpful. That's the first advisor to start with.

Once you know where you need help, think about who has already helped you understand that part of your business. It could be someone you know, someone whose work you have followed, or someone new you need to research. The name is only a starting point. Their work still has to pass the five tests.

Those five questions help you decide whether that person could be useful in the role. You aren't looking for a perfect advisor. You're checking whether their actual work matches the area where you need help and whether their way of working makes sense for your business. If you can't find enough of their own work to answer those questions, don't let AI fill in the blanks. Keep them as someone you learn from and keep looking.

If they do fit, test the advisor on one real scenario you're already working through. The test isn't only about the answer you get. It shows the perspective that advisor brings to the situation. It also shows whether that perspective factors in the way your business actually works and helps give more direction to a decision you've been waffling on.

I learned this the long way by building the full Board, putting it into real work, and then fixing the places where the advice did not fit. I turned that process into Build Your Own Advisor and Council.

Instead of trying to build the whole thing at once, it walks you through choosing one advisor, researching their work, testing the fit, and deciding whether they earn a place. Then it helps you build a small Council when a decision needs more than one perspective. It includes four guided AI prompts, three simple templates, and a completed example that you can use with the AI you already use.

You don't need a full board to find out whether this can help. Start with the part of your business where another perspective would be useful, and build one advisor from there.

If you want the full process I used, Build Your Own Advisor and Council is available when you become a paid annual subscriber to DigiNav Compass Signal.

Become a paid annual subscriber and get Build Your Own Advisor and Council.

TL;DR: Working solo with AI made it easy to move fast while staying stuck inside my own thinking. I built a virtual Board of Advisors to replace the mentors and mastermind groups I once relied on. It evolved into an Advisor for moments when I truly do not know what I think, a Council for high-impact decisions, and one important rule: most everyday work needs neither.

The voice I still reach for

All through life, you rely on other people to guide the way. In school it's teachers, counselors, and coaches. At work it's bosses, coworkers, and colleagues.

And then there was my Grams.

She was the person you went to for all the sage advice you needed while navigating this thing called life. Her opinions were trusted, and that trust came from lived experience: raising her own humans, and then their humans. Her circle from church and card club and the community trusted her and her advice too.

No matter what you told her, she listened patiently, asked the right questions, and never judged. She was the safe space. Sometimes she had all the answers. Other times she had just enough to keep you moving (and alive).

There's a saying: you don't know what you don't know. Grams was all about that. If she didn't know, she told you so, but she gave you something to think about. That's what caused the aha moments. The ones that make you think, shoot, why didn't I think of that?

I still find myself asking, what would Grams do?

So when I started building DigiNav and working with AI, I knew I needed more voices, insights, and opinions.

Hi, I’m Lee. Welcome to DigiNav Compass Signal. Around here, I share what happens when AI and technology meet a real business—the useful parts, the messy parts, and what you can take into your own work.

Subscribe now

Today, I’ll be showing you how my AI Board of Advisors went from a list of thinkers I admired to two kinds of help I actually use, and the mistakes it took to get there.

When the sounding board disappeared

For the longest time, I was part of mastermind groups, and I had mentors I could go to when I was stuck. That was my sounding board. Someone outside my own head who would listen, push back, and ask the question I hadn't thought to ask myself.

But when you work solo without a team or employees, usually with the help of AI, you seem to be on the island alone.

When working alone or even with AI, it's easy to get stuck in your own thinking. There's no mastermind call on the calendar and no mentor to catch you when an idea has been going in circles for a week. You can move fast and still be stuck.

So I thought, why not build a Board of Advisors to help move me along? I'd seen other people creating their own sounding boards of virtual experts, and I wanted more voices, insights, and perspectives in the room.

But I didn't want to collect impressive names or simulate famous people. I wanted bounded, evidence-based perspectives that fit my real business, my capacity, my goals, and the decisions actually in front of me.

✳︎ ✳︎ ✳︎

A list of experts was not yet a working board

The first version wasn't a board at all. It was a research plan built around people whose work I respected. Herbert Simon. Kahneman and Tversky. Gary Klein. I wanted to pull useful ideas from their books, papers, and lectures and turn them into decision lenses I could use.

That's a good library. It's still not a board.

The difference became clear when I put the first working panel in front of real business decisions. It was strong at checking decisions and keeping me honest. But the business coverage was uneven. No one covered the complete sales job, and the advice on monetization and audience growth leaned too heavily on a high-volume creator model that didn't fit a small, relationship-led advisory business.

So the question changed. It was no longer, “Whose ideas do I admire?” It became, “What job needs to be done, and whose documented method fits a business shaped like mine?”

You don't need to build a full board to use that shift. A few questions worth borrowing are:

• What kind of decision or work do I need help with?

• What perspective is missing from the voices I already trust?

• Does this person's method fit the way I actually work and live?

A respected name was not enough to earn a seat. The person needed a clear area of expertise, work I could verify, and experience that fit a solo, owner-led business with real limits on time, capacity, and budget. I also wanted their blind spots recorded. If everyone agreed or pushed in the same direction, the board would only give me a more convincing version of my own thinking.

Who sat on the board mattered too. When I said I wanted diversity, I wasn't asking for a demographic checklist. I wanted different kinds of lived perspective: women and men, people who understood building around children, caregiving, limited capacity, and other demands that shape how work gets done. The advice had to fit a real life, not only a business model.

The rebuilt board became seven women and four men, with different jobs and different reasons for entering a decision. They were not AI impersonations of famous people. Each profile had to stay bounded by documented work and methods, and each advisor needed to be able to disagree with me and with the others.

That gave me a working board on paper. It still didn't tell me whether I had built the right way to use it.

Then I turned it on.

✳︎ ✳︎ ✳︎

When a good-looking answer exposed the wrong design

At first the board sat inside my weekly review system, the same automated process that reviewed my daily notes, gathered the week's insights, planned the next week, and was supposed to clean up.

And the advisors, as one big panel, horned in on the results of the week. Whether I had a decision worth their time or not.

That was mistake number one. Well, two, if you count the board assembly.

I wrote about that weekly system before, so I won't repeat it here. The short version: a missing cleanup step showed me that capture, cleanup, planning, and advice were all tangled together, and the advisors needed to be their own process.

Two kinds of help, not one

Then I came across a system from another creator that split the roles of the advisors. The best way I can describe it is having a mastermind group and a board of directors.

• The mastermind helped with stuck decisions, to help me get out of my own way. They knew the one thing that was niggling in the back of my mind and pushed on it.

• The board of directors (now the Council) looked at the big picture. They could use the relevant DecisionVerse context, which is my second-brain type of setup, and they focused on growth and movement, not just what's the issue today.

So I rebuilt the Advisor and Council as two separate skills.

One of the tests: the reach plan

I had a 12-week reach plan built around a real problem: I'm having trouble promoting my articles. The goal for this quarter is to expand reach beyond Substack, so I want to start using other social media to drive traffic.

When I took the plan to the Council, ChatGPT brought in the Advisors instead. I said no. This is a bigger decision, and it needs the Council.

That was the first sign that the roles were not as clear as I thought. I corrected the route, and the Council did the big-picture work the plan needed.

Part of the plan was to take an article and expand it with YouTube videos that pointed back to the article, the perks, or whatever the promotion was. The kind of thing you could totally get on board with.

Where the problem became impossible to miss

Later, as I worked through the YouTube part of the plan, the Advisor kept showing up. It was being invoked on almost every question I asked, including ordinary questions that should have stayed inside the conversation.

It wasn't waiting until I was genuinely stuck. It was turning routine project work into a reason to bring another voice into the room.

That was a terrible way to work through a problem. The first mix-up showed me that the Council and Advisor roles needed to be separate. The repeated interruptions showed me that separation was not enough. The Advisor needed stronger guardrails for when to stay out of ordinary work.

✳︎ ✳︎ ✳︎

Not every decision needs another voice in the room

The guardrails sorted out Council versus Advisors. They didn't sort out Advisors versus everything else.

Working in projects, I kept seeing the same thing. "Let me ask the advisors. Let me pull in the advisors. Let me talk to the advisors." On ordinary questions. On routine project decisions. Once, inside a Council thread, where I'd already picked the kind of help I wanted, and I had to tell them to stop.

It's kind of like they've become a busybody of ChatGPT threads.

The advisors shouldn't be pulled in on every question and decision. But nothing in their instructions said so. Their job was written as "give direct decision support," which is broad enough to cover almost anything I do with AI. There were rules for what the Advisor couldn't execute. There was no rule for when it should stay out of the room.

Two kinds of "I don't know"

The real fix came from a thing I'd been saying without hearing it.

A lot of times I'll be working through something, get asked a question, and say I don't know. Sometimes we just get clear on it, because I didn't understand how the question was worded. But other times I really don't know how to answer it. Not because the question wasn't clear, but because I'm not clear in my mind.

Only the second one is an Advisor moment. The first one is just a conversation.

So the Advisor got an invocation boundary. Ordinary questions, routine project decisions, and normal back-and-forth stay in the current conversation. "I don't know" is not an automatic trigger. Clarify the question first. Offer the Advisor only when I still don't know what I think.

Three levels, not one

Getting there meant naming what each kind of help is for.

• Everyday work stays in the conversation. Most decisions don't need another voice in the room.

• The Advisors are for when I'm genuinely stuck. Short back-and-forth, different perspectives, help hearing my own thinking. A mastermind, not a committee.

• The Council is for the big ones. Costly, hard to reverse, uncertain, or prone to my blind spots. Two to four named advisors, positions formed independently, disagreement kept visible, and I make the call.

Either one can be skipped. Nothing has to pass through all three.

A simple version of the boundary is something you can borrow too:

Keep ordinary questions and routine project decisions in the current conversation. If I say I don’t know, clarify the question first. Bring in the Advisor only if I still don’t know what I think. Use the Council for the bigger decisions that are costly, hard to reverse, or likely to expose a blind spot. 

Where it stands

The new boundary is in place. What I can't tell you yet is whether the busybody is gone for good. That takes more ordinary conversations than I've had since the change.

I still ask what Grams would do. She listened, asked the right questions, and gave me enough to keep moving, even when that wasn't a complete answer. Turns out that was the design all along.

TL;DR: A project does not need a full cleanup before you stop. It needs a clear return point: where it stands, what remains unsettled, and the exact place to begin next time. The new paid-subscriber perk Pick Up Where I Left Off turns that rule into three prompts and three simple project files.

The question that waits at the top of every restart

My email sorter worked. That was the part I walked you through in the last article; keep the parts that worked, ditch the things that did not and give each layer a job to do. Which uncovered the missing piece - one was unassigned. Not because the process was broken or not working. But because these things happen.

And that whole email system was itself a project. The first sorting script, the test runs, the layers, the revamp - none of it happened in one sitting. It started back in March and moved in stretches, with gaps between them, until the revamp landed in August.

Every one of those gaps ends the same way any project gap does. You come back after a few days, or a few weeks, and before any real work happens, it's like you're working an archaeology dig:

• Which file is current?

• What did I decide about that?

• Where exactly did I stop?

The project didn't fall apart while you were gone. It just never got told where you'd be coming back in.

That's the unassigned job: leaving a return point. And it comes down to one rule.

Before you stop working, record the exact place where you should begin next time.

Not a summary of everything you did. Not a cleanup of the whole folder. One recorded, exact place to restart.

✳︎ ✳︎ ✳︎

✳︎ ✳︎ ✳︎

What a return point actually records

The rule only works if the return point is exact. "Keep working on the website" is not a return point. It's a reminder that a project exists, and you already knew that.

Before you stop, record three things:

• Where the project stands now. The current truth, not the history. What's done, what's decided, what's in play.

• What remains unsettled. The open questions and the decisions you haven't made yet, so future-you doesn't mistake a maybe for a done.

• The exact place to begin next time. One spot. The specific file, the specific section, the specific next move.

That's the whole rule in practice, and it takes a few minutes at the end of a work session. Not a project plan. Three short answers, written down where the project lives.

The difference shows up on the other side of the gap. You open one note and start working. The current state is there, the open questions are there, and the first move is already named.

Try it on one project. Pick something active, or something paused that you've been avoiding because restarting it feels like work before the work. At the end of your next work session, ask the AI you were just working with to record those three things. Review what it wrote, save it with the project, and walk away. That's it.

The complete method: Pick Up Where I Left Off

And there's nothing wrong with recording those three things at the end of a work session in one move, because it stands on its own. You can stop right there and be better off than you were before.

But a return point lands best when a project has a small, steady place to hold on to. And this is how Pick Up Where I Left Off was built: three prompts and three Markdown files that live in a project folder.

• Project Status.md holds the current truth and the exact return point.

• Backstory.md holds the stable context that must not be lost when you open a new AI conversation.

• Punch List.md holds only the work that remains. Finished work doesn't linger on the active list.

The three prompts match the three moments a project actually has: one sets the files up from the material you already have, one closes a work session by recording the return point, and one restarts you when you come back. Your drafts and notes stay where they are. The files stay with you, and they work with whatever AI you already use - file uploads, a local folder, or plain pasted text.

Pick Up Where I Left Off is a subscriber perk. If you're a paid subscriber, the full package - prompts, templates, and a completed example from source notes through restart - is waiting for you.

✳︎ ✳︎ ✳︎

My best friend, who I camped with for years, lives about an hour or so from me, and there are many ways to get there: a major highway, a secondary highway or a bunch of back roads.

I always take the same highway, get off at an exit, and circle back to her house, because that's just the way that I know. Even so, the GPS goes on for the drive, mostly because there is this one road to turn on, and I always overshoot it.

Then this one time there was a really bad accident that forced a detour. Somewhere along that road the thought showed up: this is a route I need to take more often. Somewhere in the back of my mind, I probably knew it was an easier drive: less traffic, no three-lane highways, just an old-school two-lane road, the kind you took before they built major intersections. It was a much more pleasant drive.

And that's when it hit me.

Sometimes we travel the same path over and over and over again, and we forget that there are alternate or better routes.

The accident forced the different route. Nothing forced me to reconsider the email system, so the comparison only goes so far. What stayed with me was how easy it is to keep using the familiar route after a better-fitting option appears.

✳︎ ✳︎ ✳︎

Hi, I’m Lee. Welcome to DigiNav Compass Signal. Around here, I share what happens when AI and technology meet a real business—the useful parts, the messy parts, and what you can take into your own work.

Subscribe now

Today, I’ll be showing you the decisions behind my email system, from the first sorting script to the two AIs that now split the work, and the question that stopped me from replacing a system that only needed a different job.

✳︎ ✳︎ ✳︎

The first job was sorting

My email had been on the same familiar route for a while too.

Back in March the project was a Chief of Staff dashboard: one place to pull information into, so the day could start with what needed attention instead of with a hunt for it. Email was the weak link. The dashboard could show how much was waiting. It couldn't show what kind of waiting.

• A client question sat next to a plugin update notification.

• Doing anything useful still meant opening Gmail and sorting through everything by hand.

• Which meant checking email to figure out whether it was time to check email.

If that sounds familiar, the next part probably will too. The wish was clear: let AI handle it. What wasn't clear was what "it" was. Email felt like one big problem, and one big problem invites you to fix everything at once.

Then one line in 's piece on the levels of AI automation cut through it: "I wanted AI to read my inbox for me." That was the problem, stated plainly. Not agents making decisions. Not workflows routing data between platforms. Just sorting.

So that was the first decision. Stop treating email as one big problem, and give the first job a name:

Sorting. Tell me whether anything in Gmail deserves my attention, so I stop opening Gmail to find out.

That decision set the limit for everything that came next. This was not a Gmail rebuild. It was one question, answered.

✳︎ ✳︎ ✳︎

What would make the sorter trustworthy?

Naming the job was the easy part. The harder decision was what it would take to believe the sorter was doing it.

Anyone who has handed a task to automation knows this moment. The thing runs. The results come back. And the question underneath is not "did it work?" but "how would you know?"

The early test runs made that question impossible to skip:

• Run one: every email landed in Action Items. The rules were not specific enough about what counted as noise, so the script defaulted to the safest bucket.

• Tighter rules: the pendulum swung. Everything landed in Read Later.

• Then the disappearing emails: the script archived before the labels fully applied, so messages ended up with no inbox and no folder.

None of that was the decision. The decision was settling, up front, on the evidence that would count. Without that, the second-guessing never ends.

The evidence was an audit trail. Every email that went through the script got tagged _Processed. When the counts didn't add up or a category felt light, that folder showed exactly what the script had done. It was the safety net I didn't know I needed: a way to refine the rules without losing anything and without going back to sorting the inbox by hand.

And even with the evidence in place, the anxiety stayed. The inbox was nearly empty, and the brain reads that as something's broken. A quick peek to see if anything got missed. Then another. The same habit the script was supposed to remove, just in a new costume.

So after a few more test runs, a look through the Processed folder confirmed the sorting was accurate, and the decision got made out loud:

Stop opening the inbox. Go to the folders. Work from there.

That sounds simple. It was the hardest part of the entire build. It was also the moment the first job counted as done: there was evidence the sorter worked, and a commitment to act like it.

✳︎ ✳︎ ✳︎

When sorting solved the wrong problem

For a while, the sorter did exactly what it was asked. The inbox stayed clean. The compulsive refreshing stopped. The first twenty minutes of the day were no longer spent sorting; they were spent deciding what to act on.

That last part is where the next decision was hiding.

One morning, coffee in hand, it was time to go through the Action folder and set the day. And a new anxiety started building. This wasn't supposed to happen anymore. The triage was dialed in, the rules were tested, and the whole thing had been written about publicly. Yet the Action folder was heavy:

• a hefty WordPress migration that needed to go on the schedule

• a two-minute yes-or-no for a different client

• three emails that had been circled for days

• a vendor follow-up promised on Tuesday

All muddled together, all the same, there was no way to tell at a glance which was the quick reply and which was going to eat several hours of the day.

And the second guessing started again. Is this actually working, or did I just move this to another mess? And underneath it all was the fear that the thing I had just told everyone was working was about to prove me wrong again.

This is the moment a lot of systems get replaced. A working tool produces a new kind of friction, and the friction gets read as failure. The easy move is to decide the sorter was wrong and go find something else.

The harder decision was to admit what was right in front of me. The sorter was doing its job. It had never been asked to do this. Sorting answers: what kind of email is this? The Action folder was asking a way different question: which of these needs me first, and how much time will it take? The question had been sitting with me the whole time. The sorter just made it visible.

Successful sorting did not solve attention. It exposed that attention was a separate job.

Admitting it was uncomfortable because that meant the system wasn't finished when I clearly said it was in the article. But it was the first time that the real gap had appeared.

✳︎ ✳︎ ✳︎

Every layer taught me something

Now that the problem was revealed, the natural move was to add something that could hold the missing context. So that is what I did, more than once. Each layer had a decision that made sense at the time, and each one taught something that the last one couldn't.

Google Tasks was the first try, because it was right there: same ecosystem, free, one click. A handful of action items went over and landed as context-free one-liners. No sender, no subject, no link back to the thread. Follow up re: site sat next to approve invoice with nothing to say which was which. They got stuck there. The lesson was quick: moving an item somewhere else does not add the context it was missing.

The Google Sheet came next, and it came out of pushing back. The first suggestions were a third-party ticket system, a separate project management app, and a form for clients to fill out. None of those fit. The email was what was working; the problem started after the email hit the Action folder, and the answer needed to live in the same place. So the Sheet got built, on purpose, with the thinking behind it rather than a generic template: who is asking, the subject, the priority, a link back to the thread, and one row per email inside a thread so nothing collapsed into a single flat line. A small script filled it in the background.

And it worked. Action items were captured with context, emails could be archived without losing the ask, and the Sheet fed straight into the morning rhythm. That was not nothing.

The Notion Attention Queue came later and made the work visible in a different way. It had views for what needed a decision, what was ready to act on, what was underway, and what was waiting. Each piece of work became a card moving through six stages from arrival to finished. It organized the work, but it was still another tracking surface.

Here is what was true about all three, and it took a while to say it plainly:

• Each one worked, technically. None of them was a broken tool.

• Each one added context or tracking. Each one also added a surface to maintain.

• None of them answered the question from the Action folder: what needs my attention now, and why?

It only added a layer of tracking, not priorities.

This became a pretty solid system. It just wasn't the system for me. And none of it was wasted. Every layer and change showed more clearly where the gap actually was. Not in the tools, but in what they were being asked to do. I knew in my gut something was off, and now I could see the results from what was actually not working.

✳︎ ✳︎ ✳︎

The job I had never assigned

At some point, what was off was the hardest to miss. Every fix had been another layer: a task list, a Sheet, a database. Each one held more context than the last, and the question to kick it off every day remained the same. The next decision wasn't about which tool to try (should I use Sheets or Notion or something else). It was to stop adding to the process and go back to the question that had been hiding the whole time.

That is harder than it sounds, but the easy diagnosis was sitting right there. It wasn't that AI wasn't smart enough, or I was using the wrong model or the wrong tools for the follow-up. I wasn't copying other people's systems. I had tried some of those. But I was trying to build a system that wasn't based on how I actually worked. When I first started building this, it wasn't about the productivity of making sure that everyone's needs were addressed and tasks were completed. It was about how it fit into my day and my energy levels throughout the day.

The honest version was less flattering.

AI was not the problem. The problem was how I defined the process.

Back in March, the job had been defined as sorting. The email triage was all about sorting. And every layer that was added after that was about tracking, but tracking and sorting are two different issues. What was never really addressed was my attention. What got addressed right away, or what got kicked down the road because I didn't have the focus to deal with it in that moment, and that's when things started to get lost. So the redefinition came down to pulling the three jobs apart:

• Sorting: what kind of email is this? Repeatable. Runs in the background.

• Tracking: what does this piece need from me? Where does it stand? It is a record, not a verdict.

• Attention: what needs me now, and why? A judgment call made with context each time.

The first two were already working, but the third had never been addressed. It had been expected to fall out of the other two, and quite frankly, it never did. That is why the system never quite worked. And that is why the gut check is so important.

If you are in the same back and forth, where the system works but something is off, here is one thing you can ask AI before you swap anything out. Paste it into whatever AI you already use and answer its questions honestly:

I have a system that works technically but does not feel right. Before I replace it, help me figure out what job I am actually asking it to do. Ask me what the system does now, what decision I still end up making by hand every day, and what I am expecting it to do that I never told it to do. Then tell me whether I have one job here or several jobs I have been treating as one. Do not recommend a new tool yet.

✳︎ ✳︎ ✳︎

Keep what works. Split the jobs.

Once the three jobs were pulled apart, the rebuild was small. Nothing got thrown out. The decision was to keep what had already proven itself and make each job's boundary clear, instead of asking one setup to do all three.

The sorting job stayed where it was working. The tested steps from the March build came forward: sort on a schedule, use a few clear labels, and keep the inbox out of it. Today that job belongs to a scheduled Claude task. It runs every two hours, every day, and it sorts into three labels: _Needs Me, _Keep, and _Get It Out of Here. It is the repeatable background job, and it does not need me in the room.

The attention job got its own surface. This is the piece that had never been assigned. It now lives in a dedicated ChatGPT Work task, the Email Brief. When I ask it to run, it reads the permitted Gmail locations, including Claude's three labels, checks the full thread and my Sent mail, and comes back with up to five things that need me, why each one matters, when to respond, and a direct link to the thread. It does not re-sort. It does not touch Gmail. It reads, ranks, and explains, and then we talk about it.

Tracking stayed inside the brief as state: New, Changed, or Carried forward. It records what remains without pretending to decide priority on its own.

This part came from an outside example. 's email process starts with a dedicated Chief of Staff thread, a single ongoing conversation that takes the place of opening Gmail. I took that one piece. Not his reply workflow, not his calendar steps, just the idea that attention is a conversation, not a list. It helped me see what my system was missing, and I adapted it to my own boundaries.

A few things worth being clear about, because this is where people tend to draw the wrong conclusion:

• This is not a model verdict. Either tool could do more than one of these jobs. I pay for both, so I am using each where it fits the job and how I work.

• Adding a second AI was not the breakthrough. Separating two jobs that had been treated as one was. The second AI is just who got the second job.

• The old work is all still in here. The labels, the schedule, the pushback on tools that pulled me out of Workspace. I pulled the best, tested process steps and improved on those.

What this looks like on a normal morning: the sorting has already happened. I open one pinned conversation, give a quick update on what I have already tackled since last time, and ask it to update the brief. Back comes a short, ranked response that divides attention into three sections:

• important emails and follow-ups

• subscriptions I need to review

• suggested filters

That covers the deadline-driven actions and responses I need to take, and it also surfaces the items that are getting in the way. It is the answer the Action folder could never give on its own.

What needs my attention now, and why.

✳︎ ✳︎ ✳︎

Review the system you already have

None of this needed a new platform. It needed a better look at the decisions that were already made, in the order they were made. If you have a system that works on paper and still leaves you with that morning feeling, the same review is available to you. It is not a rebuild. It is a walk back through your own decisions:

• What job did you give it? Say the first job out loud, the way it was defined when you built it. Mine was sorting. Yours might be capture, or scheduling, or follow-up.

• What evidence told you it was working? Not a feeling. The thing you checked. If you never decided what would count, that is where the second-guessing comes from.

• What decision still lands on you every day? That is the job the system was never given. For me it was attention: what needs me now and why.

• Which parts already earned their keep? Those stay. The labels, the schedule, the pushback that kept things inside the tools you already use. None of that is waste.

• Which single job is mismatched? Change that one. Give it its own owner, its own surface, its own question. Leave the rest alone.

That last step is what the detour taught me. The highway was never wrong. It just wasn't the only route, and I had stopped checking. The same was true of the inbox.

If you want a place to start, paste this into the AI you already use:

Walk me through a review of one system I already run. Ask me, one at a time: what job I originally gave it, what evidence shows it is doing that job, what decision I still make by hand every day, and which parts have earned their keep. Then name the one job that is mismatched and suggest the smallest change that gives that job its own owner. Keep everything else as it is.

Before you replace it, find out what you actually asked it to do

Yesterday, I shared why I decided to move my subscription tracking from Notion to Google Sheets.

The decision was already made. What happened after that is the useful part.

This is a look at the prompts, decisions, corrections, and checkpoints that moved the work from an idea to a working system.

The decision was already made: subscription tracking was moving from Notion to Google Sheets.

What happened after that is the useful part.

This is not a tutorial for copying my setup. Your tools, data, and rules may be different. It is a look at how the work moved from an idea to a working system, including the prompts that moved it forward and the places where human judgment still mattered.

If you have information stuck in a tool that no longer does what you need, there may be something here worth borrowing.

The working sequence

Before anything was moved, the project was divided into these steps of what needs to happen:

1. What does the finished system need to do?

2. What is actually in the current system?

3. Which rules require a human decision?

4. How does the plan move from Chat into Work?

5. Can the new system work correctly before automation?

6. How will actual transaction data reach the Sheet?

7. What can be automated safely, and what should stay visible?

8. Move the lifetime deals into a separate tracking area

The sequence kept “build the automation” from becoming one large, vague request.

✳︎ ✳︎ ✳︎

Step 1: Start with the job, not the replacement tool

The first prompt gave ChatGPT the destination and the reason for the move:

What I asked:
“Let’s come up with a five-step game plan to convert this from a Notion database to a Google Sheet with automation.”

A version you can adapt:
I want to move [current system] from [current tool] to [new tool] because [reason]. Before changing anything, create a phased plan that covers inspection, design, migration, validation, and automation. Show where you will need my judgment or access.

The end state was also clear. The Sheet should connect with Chase transactions, identify paid subscriptions, and show changes such as a price increase.

ChatGPT proposed this order:

1. Audit the Notion database.

2. Design the Sheet around expected versus actual charges.

3. Move and clean the data.

4. Build the Chase matching process.

5. Add the monitoring layer.

The useful part was not the number of steps. It was the order. The Sheet needed to “tell the truth manually” before Chase became an evidence source.

✳︎ ✳︎ ✳︎

The plan was the easy part.

The real work started when ChatGPT inspected the existing database, found the contradictions, and brought the decisions back to me.

The rest of this article includes the prompts, corrections, checkpoints, and automation choices that moved the project from a plan to a working system.

Read more

It was easy to master walking on them, once you were up. And that was the problem.

The easiest way was to have someone help you. They held onto the stilts. You hoisted yourself up, and off you went, but there were times, more often than not, when nobody was around to help. That's where the Back Porch Steps came in. Sure, they were a little bit hard to navigate at first, and I can say not at all that safe, but they gave you the boost you needed to get started. And the same thing happened the other day when I found out ChatGPT could give you the same kind of boost.

The handoff I always did myself

It all started in ChatGPT that day in a regular chat, and that was on purpose. I was thinking something through, and thinking is not something I want to spend my time and tokens on in Codex Work. Work is where the building happens, you know, the doing part. So I stay in chat while it's still a conversation, mapping and planning, and when it turns into something real, a punch list or a plan, I move the project over to Codex Work.

What I happened to be working out that day was my subscription tracking migration:

• Get it out of Notion and into Google Sheets.

• Link my Chase statements to the spreadsheet.

• Let it record the payments and what was due when.

But it wasn't the subject that mattered. What mattered was that it started as a conversation in Chat and it needed to end up as an action plan in Codex.

So once it was all sorted out, I asked for what I always ask for:

Give me a recap, a punch list, and context of this conversation in Markdown files so that I can move it over to work.

And I already know what comes next. Download the files into the right folder, move it over, and start the conversation again with the proper assets and context. All of that just to move something already finished into another space without having to re-type or rehash the whole conversation.

✳︎ ✳︎ ✳︎

Continue to work?

Except that isn't what happened. I asked for the files, and instead of handing me the markdown files to go carry somewhere else, I got this:

Continue With Work?

You want me to create and move files into your local Obsidian vault, so Work mode is the right place to perform those file operations safely.

And then a button. Continue with Work.

To be fair, I had told it where the files were going. That was in the ask. What I didn't expect was for it to treat that as the next step instead of the end of the conversation.

So of course I clicked Continue with Work. And it went and opened up work and continued the conversation there. It updated the project and added the three files I had asked it to create. It actually found the open folder in Obsidian and did the work from there.

No more download, move the files, reorient the conversation. The part that I always did manually was now happening for me. Automatically.

✳︎ ✳︎ ✳︎

Nothing had to be explained twice

My reaction to finding this new feature was a holy crap moment. Is this a game changer for using the thinking side and the doing side? The best part was the time and energy it saves. I mean, nobody minds skipping a few steps.

But here's what happens when one hand works with the other. I didn't have to retype any of the planning and clarification that I got as part of the setup, because chat knew that I wanted three files:

• the recap of what we had worked out

• the punch list of what still needed to happen

• the context of why we made the decisions that we made

So while a lot of the heavy lifting was done with the regular AI chat, it carried over word for word, thought to thought, and clarification to clarification. The first thing that happened by moving from one model to the next wasn't a reorientation. It was picking up the punch list and starting to work on it, because at this point I knew the rest of the build could be done by ChatGPT work and not me manually making the changes.

It was ready to pull the subscriptions out of Notion and put them into Sheets, and here is how that went:

• First it set up the Google Sheet based on the decisions that we had made on what to keep and what to dump.

• Then we worked through connecting the Chase statement, because there wasn't a connection or an MCP available, but there was definitely a workaround.

• That let me bring in the reconciliation and the daily automation so that things could stay current without me having to lift a finger to do it.

So in the end it wasn't about starting from square one. It was that nothing had to be explained twice, and we were automatically able to move from one phase to the next.

✳︎ ✳︎ ✳︎

I'm doing more

Here's the thing though. I wasn't always this careful about where the work happened. For a long time I didn't separate chat and work at all, because I pay for the higher level and figured it was all one big pool. Why would I care which side of the house I was standing in?

What changed wasn't the tools. I'm doing more.

• Building and automating instead of just planning to.

• Several projects moving at the same time.

• Images and design work running through other tools.

• Research and plan review that has to happen before any of it is worth building.

And when you're doing that much, one big pool stops being a pool. I found that out the day I hit the limit in work and had to burn one of my resets just to keep going. That's not the point of the story, but it's what made me actually look at where it was all going.

What I found was that I'd been spending build capacity on conversation. Thinking out loud in the place that was supposed to be doing the work, and then wondering why there wasn't enough left when it was time to build. The old way wasn't wrong when I started. It just got too blunt for the amount of work moving through it now.

It's about working smarter.

✳︎ ✳︎ ✳︎

Where I've landed, for now

So here's where I've landed, at least for now.

• Chat gets the thinking. Planning, brainstorming, mapping something out, the general questions I need answered before I even know what I'm building.

• Work gets the doing. Acting for me, saving the notes, scheduling the tasks, running the complex steps once the goal and the punch list are clear.

• Everything else gets the parts it's better at. Research, plan review, the image and design work. Claude and the others are in that mix too.

The other thing that changed is how I open. Now I start every conversation with some version of this:

Today we're working on [whatever it is] and the goal is [whatever the goal is]. Let's make sure we're on the same page. Ask me clarifying questions until we both understand what needs to happen to get there. goes here

It sounds like a small thing, but it front-loads the arguing. We sort out what I actually mean while it's still cheap to sort out, and by the time we're done I either have a punch list or I know I'm not ready for one.

What tells me it's time to move isn't a number of messages or how long I've been sitting in a conversation. It's clarity. When I can say what the goal is and what the steps are, it's ready to go get done. When I can't, I'm not finished thinking yet and it stays right where it is.

And I want to be careful here, because this is what works for me, right now, with the work I'm doing and the tools I'm paying for. Yours might split somewhere else entirely. The products keep changing, what they'll do for you keeps changing, and my own setup will change again too.

It's not about the ai but knowing the best practices for you.

✳︎ ✳︎ ✳︎

The part that's easy to miss

By the end of that day, the whole thing was finished. The subscription tracker was migrated from Notion to Sheets. We added in the automation layer that pulls the data in from Chase for the reconciliation, and now it's keeping current on its own.

I have different views set up so that I can see what's renewing without going and digging for it. I even have the entire Lifetime Deals inventory back added to the information, which I honestly was going to give up on because that was a big undertaking itself.

And this is the beauty of finding the quiet parts.

Just like me walking on those stilts, there isn't one clear way to make something happen, but you need to find the best way for you to move forward. In this case, it was starting a conversation in chat, but letting work do the implementation for me. Same difference of having someone help you get up and moving or using the step that takes you to the next phase.

And here's what I keep coming back to. That change happened behind the scenes. Nobody announced it, nothing told me that the part I'd been doing by hand was now just handled. I only found it because of the way I work now, thinking in one place and building in the other. If I had stayed put in one of them, that button never would have come up.

Which makes me wonder what else has changed while I've been busy doing things the way I've always done them. That's usually where the difference is, and it's the easiest thing in the world to miss.

That was the whole thing.

Now you have to figure out which remote actually turns the TV on, which one gets you to the right channel or input, and (if you’re in my house) which one runs the sound bar. Three remotes to do one thing: watch something.

I know which remote is which just by the look and the shape of it. That doesn’t stop me from grabbing the wrong one. Constantly.

I mean, if it were up to me you wouldn’t need a remote at all, because the TV would never be on. But that’s not the reality of the house I live in.

Which is why universal remotes exist. Not because I need to understand how they are all wired together in this one remote. I just want to press one button and have it work like it should.

The same thing had been happening with some of the AI processes I was building.

Start the research in ChatGPT. Create the images in Midjourney. Edit and revamp in Claude.

Every one of those is the right tool for that job. I picked them on purpose. But you end up with pieces of the process here, there, and everywhere.

And the friction isn’t that I’m using too many AI tools. It’s that I have to remember where each piece lives and which one I’m supposed to reach for next. Same as the remote basket. I know exactly which one is which, and I’m still standing there holding the wrong one.

The short version: the process now lives in one folder instead of inside any one AI. I change the instructions in one place, and whichever AI I open that morning reads the change. The finished work lands in that same folder, so my dashboard always knows where to look.

I asked for one shared folder and got two AI skills to maintain

The one that “broke the camel’s back” happened in Prepare My Day.

It’s my morning process. What am I actually moving forward today, what needs tending, what happens if there’s room. I wanted to take it out of one AI’s skills and put it into a working folder that Claude, ChatGPT, Codex, or whatever I’m using next could all work from. I said that in the request. Shared by both, or any other AI.

What came back was anything but what I asked for. A folder in Obsidian that held the master copy of the process but it added two more steps:

• Save a packaged copy as a skill in Claude.

• Upload another copy as a skill in ChatGPT.

And from there on: edit the folder, then refresh both installed copies whenever a change matters.

This is not my original ask. I dont understand.

That’s what I typed back. Exactly that, typo and all.

To which I got an explanation wrapped in a recipe analogy. The Obsidian folder is the master recipe. Claude and ChatGPT can’t read the folder on their own so they each need their own directions on their own page. Two one-time setup steps, one per app. Like having two separate recipes in the box. But the ingredients don’t change.

It was a good explanation. Clear in understanding but not exactly what I asked for.

I still don’t understand. I specifically said I don’t want it as a skill that’s installed in each AI because it needs to be AI agnostic, and you’re telling me to save it as a skill.

Because then I have two skills that I have to keep updated if I make any changes.

That was the actual objection. Not the explanation. The arrangement.

And I already had the model I meant sitting right there. Content Lab works from a folder. The WPSS article writing works from a folder. The mentors and the advisors work from a folder. I open the folder, the AI reads what’s in it, and when I change something, I change it once.

I wasn’t confused about how to maintain the copies. I was asking not to have copies to maintain.

Sharing the instructions wasn’t enough. The results needed one place too

So I took the folder to Codex and asked a simpler question. Is this actually set up to work for another model?

Not quite. Everything in the folder was usable. It was just still built to be installed, which was the exact thing I was trying to get rid of. Take that off, don’t put copies anywhere, and the folder becomes the thing every AI works from.

So I typed back three words. fix it please

And that was that. The folder finally became the thing I asked for the first time. One set of instructions, one place to change them, nothing to install anywhere.

Here’s what I hadn’t thought about at all.

Prepare My Day has two halves. The first half is the instructions. How to look at the week, what to weigh, how to land on what actually matters today. The second half is where the answer goes. Every morning that process ends by putting three cards on my Personal Dashboard, a small thing running on my Mac. What I’m moving forward. What needs tending. What happens if there’s room.

I had spent all of this energy on the first half. I never once thought about the second half as part of the thing being shared. It was just where the work ended up.

Then one morning the cards were still sitting on August 4 and it was August 5.

Moving off the skill had broken a small piece of the dashboard. It was reading from the wrong thing. Between me and Codex we found two problems. The local server was down, which explained one of them, and there was a small glitch in the call, which for AI is an easy fix. There was only one place to go looking for either of them, which I didn’t appreciate at the time. Both got sorted. The day got published and I could see the three cards again.

Fine. Done. That’s what I thought.

Then it happened again the next morning.

And this time what came back were two ways to finish it myself. Open Chrome on my Mac and say retry. Or copy a line of code into Terminal and run it.

I’ll paste something into Terminal if I have to. That isn’t the point. The point is that to make it work it needed more manual input steps. Something waiting for me every morning.

The dashboard lives on my Mac at an address only my Mac can reach. A model working in the cloud can read a shared folder just fine. It cannot reach into my machine and set something down there.

Which meant the second half had never been shared at all. It only ever worked because I was sitting at the machine where it lands.

I had shared the instructions. I missed giving the finished work the right place to land.

What an AI-agnostic workflow actually needs

Here’s where my thinking actually changed.

This was never about wanting one AI to do everything. It was about one process. Prepare My Day. I wanted to run it in whichever AI I happened to be sitting in that morning.

Right now it’s Claude. But next week it may be Codex. Next year it might be something that doesn’t exist yet.

The answer was not to give them all their own copies. The answer to the real problem was simplicity and cohesion. Not let’s match copies every time there is a change or adjustment. Because now I have to figure out where are all the places to make those changes. Instead it’s about one place, a single source of truth so that each AI is not quietly doing their own thing.

The two things that have to be shared

What I actually wanted was smaller than that:

• One live place for the instructions. I change them once, and whatever I open that morning reads the change.

• One shared place for the result to land. Whichever AI ran it, the answer goes to the same spot, and my dashboard knows where to look.

That’s the whole ask.

The tools stay different. That’s fine, and that was never the part I was trying to fix. The part I was trying to fix is that there’s one process, and whichever tool I pick should just be able to run it. Picking a different one shouldn’t mean installing something, setting something up, or keeping another copy current.

Finding the glitch in the second part was the icing on the cake. Sure it broke with the first (and second) run but it would have been harder to diagnose and fix if each AI had its own skill.

What the shared folder changed, and what still isn’t automatic

Here’s what my morning looks like now.

I open whichever AI I’m working in. I say prepare my day. It reads the process out of the folder, we go through it, and when I approve the three things, it writes them down in that same folder. The dashboard picks them up on its own. I don’t open Chrome. I don’t paste anything into Terminal. Nothing is waiting for me.

That’s the whole difference.

It took me a minute to believe it, though.

I wasn’t satisfied with taking “it’s working” for the final answer. You need to test, and a few times. Mine was several tests over different checkpoints including running from the cloud instead of the desktop. Then reviewing the actual dashboard to see three cards sitting there on my screen.

Because that was the ultimate pass/fail. Three cards, no blanks.

It’s not an automatic process though. It still needs me to kick it off. But now I choose my AI of choice, connect to the folder and say “prepare my day”. And this was the goal. One process, any AI with the same results.

The prompt to check your own workflow

And now it’s a simple ask for any other process I want to run the same way. Here it is, if you want to point it at yours.

Look at how this workflow is set up. If I change the instructions in one place, will every AI use the updated version, or will I have to update it somewhere else too? Also check where each AI saves the result. Explain anything that is being duplicated and suggest the simplest way for every AI to work from the same folder. Don't change anything until I approve it.

Back to one button: what happens when the tools change

This is one process that I took from scattered to streamlined and I have many others still.

The goal is to take advantage of whichever tech or AI is my tool of choice. Prepare My Day no longer belongs to any one AI. It isn’t sitting inside Claude, or inside Codex. It lives in a folder, and whichever one I open that morning goes and reads it.

Which brings me back to the remotes, because that was the problem I wanted to solve. Not fewer devices. The television, the sound bar, and whatever is back there are all still doing their own separate jobs. I just wanted to press one button and have it work like it should.

Because at the end of the day tech changes. AIs are advancing. The options are increasing. And it’s not which should I choose that is the key decision, but will it break what I have.

And this exercise is where you know something isn’t working, but you can’t quite put your finger on it. For me it was why can’t I use ChatGPT instead of Claude to make this work. For you it may be something else.

I don’t usually turn this on myself, but here we are.

Lately I’ve been refining the Content Lab — my multi-step process for working with AI on these articles, built to lean on my strengths and help with my weaknesses. But something is off, and so I put together some field notes.

What I wouldn’t do to be a kid again. Not because sixty isn’t nifty (how cliché), but because you just did things and didn’t overthink for perfection. Reckless abandon was my childhood. There were no refined edges. Eat the sugar and run off the calories. Jump the bike ramp without thinking about the broken bones. Wear the old, crusty sweatshirt, not caring what anyone else thinks.

We all had those moments. Some more curiosity-driven than others.

And there was nothing wrong with any of it. My edges don’t need to be refined to perfection. My spirit animal is the mess that comes with everything I do. I say what I say, and while sometimes it comes out a little wonky, people who know me get that it’s just me. It’s how I think and act and learn.

Which is exactly what the Content Lab started smoothing away — that’s the “offness” I keep coming back to. I start every session with a brain dump — the good, the bad, and the indifferent. Throw-spaghetti-at-the-wall moments, and AI organizes the information, runs an interview, and pulls out what’s missing. Tells me if there are too many directions. Helps me work out the kinks.

And it works. Or that’s what I thought. Because I looked at the numbers and the data to confirm what my gut has been telling me for far too long. Something is off. The favorite sweatshirt is a little too stiff and ironed out. And so were these articles.

Today, I’ll be showing you what came out of looking closer at a process I built to help me refine my own writing, what was actually underneath the feeling I kept pushing aside, and what that might help you notice in your own work with refining processes before you decide everything is fine.

✳︎ ✳︎ ✳︎

The guardrails were there for good reasons

The Content Lab, a multi-step process for helping me with articles, was never meant to replace me or my voice. It was built to work with me.

And over time, as I kept refining it, I added guardrails to each of the phases. They were supposed to:

• Make sure the story had a flow and gave it a real beginning and end instead of letting it trail off on a tangent.

• Add something actionable when a piece called for it.

• Stop me from overthinking or overanalyzing a decision I’d already made.

• Keep me from repeating myself once I’d already made the case.

• Give me a simple, refined message.

That was the bottom line: a nice, neat, tidy, refined message. Every one of those reasons still holds true. I wasn’t trying to sound like everybody else, or have AI write for me, or even hide behind it. I didn’t want to produce something generic, but I also didn’t want slop. I was trying to fix real weaknesses of mine:

• The tendency to circle through a thought six times before it actually landed.

• The drafts that went nowhere because there was never really an ending.

What I didn’t build in was a way to protect the parts of the process that gave me the real big picture of what I was working through.

The numbers made the feeling harder to ignore

A feeling by itself is easy to talk yourself out of, especially when the thing sitting in front of you technically works. What changed, though, was the one thing that never lies: the numbers.

of GenAI Unplugged and the Subflow connection into Substack GenAI Unplugged and the Subflow connection into Substack made surfacing those numbers easy enough to find. I used Claude to pull the information, have the conversation, and then sit with it — noodle through where the numbers were actually coming from instead of just looking at a dashboard of facts and figures. What I found, comparing my last four posts against a stretch in the spring:

• Views were down by about 25% compared to the spring range.

• Open rate was down to 18–20% recently, versus 24–25% in March and April.

• The top performers in the batch weren’t the polished pieces. They were the practical, work-driven posts about things I’d actually built, messy middle parts included.

Now, I need to be honest about what this does and doesn’t prove. New subscribers came in over that stretch, and new subscribers can quietly pull an average open rate down without anyone disengaging from anything. Topic, timing, and who’s actually in the audience right now can be doing some of that work too — and this is only from subscriber information. The numbers didn’t hand me the cause of the discomfort.

The numbers showed something was off, and that was enough to stop me from explaining the feeling away.

There was relief in it too — not because I liked what I was seeing, but because the numbers were starting to make sense of something I’d been sitting with for a while. It was a justification of the feeling, not in the “am I right or wrong” sense, but the sense that something wasn’t working, and the numbers proved it.

The relationship being edited out

Now, these numbers didn’t tell me what was actually missing. Going back and rereading the stronger pieces next to the recent ones did. The early articles were more me — messy in a very specific way. They kept:

• the wrong turns

• the questions I asked two, three, four times because the first answer didn’t hold

• the thing I got wrong before I got it closer to right

The newer pieces kept the conclusion. They just didn’t keep the walk to get me there. The messiness, what I did wrong, not just the perfect tied-in-a-bow end piece, was being taken out of the equation.

The relationship between me showing the whole process and you watching it happen was what was being edited out.

That’s the unfinished thinking that I was struggling with and that I noticed many of you struggle with. AI can summarize what we’ve learned. It’s very good at that. What it doesn’t do on its own is show how the learning happened. That’s the part that had quietly disappeared.

Not every rough edge was a flaw

For a while, I thought it was the “voice.” It was the word everyone reaches for when something written with AI feels off. But even that seemed too vague to actually diagnose anything, so I needed to sit with it a little longer.

To me, refinement is smoothing away the rough edges. I think I shot myself in the foot a little, because refinement is not my natural state. Those edges are what keep my curiosity flowing.

It’s that never-ending feeling of not knowing what I don’t know, and somehow it feels like my favorite sweatshirt on a cool summer evening — comfortable, broken in, not something I want to try to fix or replace.

It’s like when you were a kid. You didn’t ask “will this work?” — you just did it. The scraped knees, the broken arms, the knocks to the ego. But also the “hey, that was cool, we should do that again” moments — the things that actually happened and turned out to be kind of cool anyway.

Not every rough edge in my writing was a problem I wanted the Content Lab to fix.

• Worth smoothing: confusion, repetition that didn’t serve anything, missing context that made a piece hard to follow.

• Not worth smoothing: the uncertainty itself, the reversals, the place where you’re still figuring something out.

That second kind isn’t a flaw sitting in the draft or brain dump. That’s the part that lets you know someone else feels the same thing.

I don’t have a clean name for the line between those two yet, and I’m not sure I even want to figure one out. Naming it too early is how you end up with a rule that flattens the very thing you’re trying to protect.

I know where to start. I don't know the fix yet.

What I decided was to go back through every guardrail I added to the Content Lab, phase by phase, because something in there is where I think I’m getting edited out of my own process. I’m not sure that’s actually the case. It might be that the guardrails themselves got over-refined. It might be that I’ve been defining refinement wrong the whole time. It might be something else in the workflow that I haven’t quite isolated yet.

Here’s what I’m not doing:

• Stripping every guardrail out at once.

• Bolting on some new rule that tells the process to sound more “authentic” or more like me.

• Rebuilding the whole workflow off one signal before I actually know what’s causing it.

What I am doing is using this article as the test:

• Starting from the real material of the week I document, instead of a topic that’s randomly picked.

• Keeping the raw conversation intact instead of summarizing it way too early.

• Building a record of what actually happened before I let anything interpret it.

• Carrying the exact words and uncertainty forward instead of smoothing them out before the first draft.

• Watching, honestly, whether the finished piece still holds the process I went through, or whether it quietly disappears by the time you’re reading it.

Right now I have a starting point. I don’t have a reworked or repaired Content Lab. I don’t know yet which guardrail is responsible, or how much of what I saw in the numbers is even about all of this.

And if you’re running into something similar — your gut telling you something’s off — start with the numbers. Pick one part of it. You can even start like I did:

Something in my gut has been telling me [this part of my work] is off, and now I’ve got some numbers or evidence that back that feeling up: [describe what you found]. I don’t want to tear the whole thing down. Where would you start figuring out what’s actually going on?

Refining should make the work more mine

Something is off. That’s still true. Not because the Content Lab failed — it did exactly what I built it to do. What it revealed is that my idea of “working” only ever asked one question: did this get clearer? It never asked the other one: did I stay inside it?

I’m not walking away from refinement. I still want the flow, the ending, the restraint. I’m just changing what I check for before I call the process working — clearer isn’t the whole test anymore. Whether I’m still recognizably in it is the other half.

Somewhere in all of this is still the kid who’d rather create the new next ending than accept the one the story already gave her. I’d rather leave that part a little unfinished than iron it into something tidier than it actually is right now.

What did your process improve — and what might it have quietly smoothed away?

For a while I was running three homes at once: the house we live in, the RV, and wherever my husband and I landed for the winter.

Each one had its own lists, and most were good. The RV opening-and-closing list, which is really the do-not-mess-this-up list. The household list. The winter packing-and-living-away list.

Good lists. Accurate lists. And things still slipped. Not because anything on them was wrong, but because nothing coordinated them.

I'd stand in the kitchen and think, where did I keep that? Not because it was never written down. It almost always was.

What was missing was the handoff between them—a way to see what mattered for the move we were making.

You may have one home and still recognize the business version: email, the calendar, a notes app, two or three project boards, a documents folder, a handful of AI threads you keep meaning to reopen. Every piece can be accurate, and still no single place shows you where the whole thing stands right now.

The same pattern showed up later across two brands and the AI tools I use regularly. Same shape, higher stakes.

✳︎ ✳︎ ✳︎

If you read DigiNav Compass Signal, you’ll stop second-guessing your AI and automation decisions because you’ll have a way to tell what actually fits your business from what’s just noise. The expertise is already yours. You just needed a framework to trust it again.

Subscribe now

✳︎ ✳︎ ✳︎

The business version of three homes

The Decision Toolkit had its own version of the problem. The material was all there: articles, experiment notes, worksheets, prompts, drafts, a string of AI conversations. Nothing was lost. Everything was saved. And starting a work session still meant digging.

A project can be that well documented and still be hard to pick back up.

Your version might not be a toolkit. It might be an offer, a course, a client process, a content series, some internal tool you keep meaning to finish. Plenty of material behind it, and no obvious answer to "where does this actually stand today?"

That can look like:

• An answer you know you got right, sitting three days deep in an AI thread you can't re-find.

• Two or three drafts of the same piece, with no marker for which one is current.

• A card on your dashboard that tells you the project exists, but not where to pick it back up.

Every one of those is saved work. And saving something is not the same as being able to pick it back up without rebuilding it in your head first.

That rebuild is the part that quietly eats the session. Too much of the time I had set aside for the Toolkit went to reconstruction before anything new got made. The project didn't need another beginning. It needed a way back in.

✳︎ ✳︎ ✳︎

The work before the work

Come back to a project like that and the same questions run first, before a word gets written or a single decision gets made:

• Where did I keep this?

• What did we decide?

• Which version are we working on?

• What started this thread?

• What was the next thing I meant to do?

That's the work before the work. On a good day it doesn't even register. It just feels like starting slow.

When it happens often enough, it's easy to decide the problem is you. Maybe you're being inconsistent, or scattered, or bad at follow-through.

But look at what's actually happening. The time set aside to move the project forward goes to reconstruction instead. The session turns into a search party. The friction is in the project, not your character.

An AI collaborator can hit the same wall. It can pick up a thread quickly, but first it needs to know where things stand, what got decided, and which version is current—the same questions again.

Point it at the current state and it continues the project. Leave it guessing and it fills the gaps with its best version of the story, which might not match where things actually are anymore.

Then Kim Doyal, who had offered to put the business through its paces, sent back a Growth Map with one clear instruction: gather what already existed and get it in front of people.

The second part depended on the first. The material was scattered across too many places to pull together on demand. Before the Toolkit could go anywhere, it needed one place to live.

✳︎ ✳︎ ✳︎

One shared home, not one giant list

The fix wasn't a bigger folder. It was giving the Decision Toolkit one place where I could see where it actually stood without reopening its entire history.

The chapters and worksheets live there. So do the AI prompts, the punch list, the decisions we've already made, and a short project status that tells me where to pick back up. When something changes, the current files change with it.

That means I can read through the project myself or ask an AI collaborator for an update based on those files. I'm no longer hoping the right answer is hiding in the right conversation—or reopening a decision because neither of us can find where it landed.

This is closer to what finally worked with our three homes. We didn't stuff the RV list, the household list, and the winter list into one enormous master list. They remained useful for different reasons. What we needed was one place to get our bearings.

A project works the same way. Notes can remain notes. Drafts can remain drafts. The useful part is knowing which ones represent the project now, what has already been decided, and where the work resumes.

For the Toolkit, that was enough to change the work. The written material and AI prompts are complete. Instead of reconstructing those pieces again, I can use them as the foundation for the next stage: a live online version alongside the PDF and AI-guided options.

But giving a project one home only solves part of the problem. Put it down for two weeks and it can still disappear into the noise. It also needs a way to point you back.

✳︎ ✳︎ ✳︎

Make one project resumable

For the Toolkit, the piece that points me back is a short Project Status. It records where things stand and the exact spot to resume. One card on the dashboard points me back to where I left off. Quick reorientation to keep things moving.

You don't need to build any of that to get the same benefit. You can test one project by hand, without redesigning your whole system.

Pick a project that's stalled. Sometimes the problem isn't motivation. If getting oriented costs nearly as much as the next step, the return path may be what needs attention.

Try this with one stalled project. Answer these five questions:

1. Where does the current project live?

2. What has already been decided?

3. Which material is current?

4. Where exactly should the work resume?

5. What's the next move that takes it closer to another person, a customer, or a real-world test?

Want AI to help you get your bearings? Give it the current project files and ask:

Based only on these files, tell me what is current, what has already been decided, what is still open, and where I should pick the work back up. If the files don't answer something, tell me instead of guessing.

If all five have quick answers, orientation probably isn't the main thing stopping you, and that's worth knowing on its own.

If one answer is missing, start there. Repair that part of the return path before you add another tool, another dashboard card, or another AI conversation.

And you don't have to gather everything just because it's related. Keep what helps the project answer those five questions. The rest can stay where it is, available if you need it, without competing to be the current version.

A trustworthy way back in

We don't have three homes anymore. We sold the RV and moved, so three sets of lists became two, each with different needs. What stayed useful was the coordination: each place could keep its own list, while I had one place to get my bearings before shifting from one to the other.

DecisionVerse, the shared home I built for that work, is doing the same job now across two brands, a handful of active projects, and the AI collaborators I work with.

Yours might be far simpler. A single project, a couple of notes, one shared doc. The setup was never the point. The test is whether you—and anyone or any AI helping you—can see where the project stands and keep going from there.

So a stalled project may not need more discipline, and it may not need a fresh start. It may just need a trustworthy way back in.

Which of your projects would get easier to finish if you didn't have to rebuild it every time you came back?

✳︎ ✳︎ ✳︎

Right now I’m running a small beta of the Decision Toolkit — a handful of people testing it on one real decision and telling me what actually happened. It’s open to subscribers. If you’d like a spot, subscribing is the first step.

Subscribe now

And yet the next useful move kept getting harder to see, not easier. Which raised a question I didn’t particularly want to ask:

If all of this access is supposed to make things clearer, why doesn’t it?

A good cheat sheet—and the way each of us needs information to work

I was not always the best student. Not because I couldn’t do the work, but because my mind was usually too busy daydreaming to pay attention to the details of a lecture, or to read a book cover to cover just because the next test would cover it.

What I loved were crib notes. CliffsNotes. Study guides. A good cheat sheet.

I know how that sounds. But I wasn’t trying to get out of the work. I wanted what mattered where I could see it. Give me one page that showed the point, and I could do something with it. Hand me the textbook and three weeks of lecture notes, and everything important dissolved into everything else.

It took me years to understand it wasn’t always me. It wasn’t a character flaw. It was information about how my brain works.

Maybe you did fine with the textbook. Plenty of people do. But I’d bet there’s a version of this in your business right now:

• The course you bought because it worked wonders for the person who made it.

• The planning template you set up in a weekend that you quietly stopped opening by week three.

• The morning routine that fits a person whose mornings look nothing like yours.

None of that failed because you’re undisciplined. It failed the way the textbook failed me. It was built around how someone else needs information to work.

Worth noticing this week: the last time you actually had to decide something, what did you reach for? Not what you’re supposed to reach for. What you reached for.

That’s your cheat sheet. That’s the shape your information needs to be in.

I didn’t connect any of this to my business until Kim handed me a report.

✳︎ ✳︎ ✳︎

If you read DigiNav Compass Signal, you’ll stop second-guessing your AI and automation decisions because you’ll have a way to tell what actually fits your business from what’s just noise.

Subscribe now

The expertise is already yours. You just needed a way to trust it again.

✳︎ ✳︎ ✳︎

The report that moved something—and the pile most of us already have

Last week, put together her Growth Map report for me. She’d dug through my offers, my messaging, my half-built pieces, and there was real analysis behind it. But that’s not what made it useful. What made it useful was the order. Do this first, for this reason. Then this.

A study guide for my own business.

And I moved through it:

• The consulting sessions now have clear pricing.

• The messaging finally says how the two offers fit together instead of leaving people to guess.

• And the Compass Check Toolkit stopped waiting for perfect and started moving toward real testers.

What got my attention was this: not one of those changes came from new information. I already knew all of it. It was sitting in my notes, in old AI conversations, in my own gut. What I was missing was someone saying this matters now, and this can wait.

You have a pile like that too:

• The saved posts.

• The notes app with four versions of the same idea.

• The AI conversation from March where you worked out exactly how to fix your pricing, and then closed the tab.

If someone read through it all, most of what you need is probably already in there.

So the next time you catch yourself researching something you’ve already researched, try a different question.

Instead of: What else do I need to know?

Ask: What do I already know that I haven’t acted on?

For me, acting on the report felt easy. Almost suspiciously easy. And that’s what made the rest of my system look bad.

✳︎ ✳︎ ✳︎

The system I stopped using—and the one you might be maintaining out of habit

I had built myself a daily system called Daily Rhythm. It was designed around my energy instead of a schedule, because that’s how I actually work. It was supposed to be the anti-task-list. Slowly, quietly, it became a task list anyway. One I got very good at ignoring.

So I ran an accidental experiment. I stopped using it. Nothing dramatic. I just didn’t open it one day, and then I didn’t open it the next.

A week went by.

Nothing broke. I didn’t miss it. Client work happened. The pricing changes happened. My days moved fine without it.

That stung a little. I designed the thing. I built it, refined it, maintained it. And the most useful thing it ever told me was what happened when it went silent: I had been updating it more than using it.

The school version of me would have spotted it right away. My system had become the textbook, the lecture, the handouts, and the notes all at once. Everything documented. Nothing distilled.

You may have one of these:

• The weekly review you keep rescheduling.

• The project tracker that’s always one week out of date.

• The journal template with twelve prompts you answer with one word each.

Maintaining it feels responsible. Skipping it feels like a small failure. But notice what actually happens on the weeks you skip it.

If you want to know for sure, borrow my accident and make it on purpose:

1. Pick one recurring ritual.

2. Stop it for a week.

3. Pay attention to what’s genuinely missing. Not what you feel guilty about. What’s missing.

If the answer is nothing, that ritual was maintenance, not support.

Mine never came back. But stopping it opened a messier question: why had I built so many places for everything in the first place?

✳︎ ✳︎ ✳︎

I kept giving everything a place because I hadn’t decided what still deserved one

When I finally looked, the answer wasn’t complicated. The same project was living in three places. A note in Obsidian, a card in Notion, a folder on the hard drive. Not because any of those homes was wrong. Each one made sense when I chose it. But I’d never decided which one was in charge.

And it wasn’t just storage:

• Half-finished experiments sat everywhere, all still technically active.

• Answers I’d worked out completely in an AI conversation months ago were still buried in a thread somewhere.

• Projects that hadn’t earned a minute of attention in weeks kept their active status because parking them would have required deciding about them.

My first instinct was the obvious one: design a better system. I asked Claude for help, and it enthusiastically handed me a five-move replacement process with review cycles. I’d asked for simpler and received a new department.

So I took the whole thing into ChatGPT and made it keep asking me questions until I could say what I actually wanted. Not another planning process. Not a smarter to-do list. Help seeing what mattered now.

Somewhere in that back-and-forth, the real problem showed up in one sentence:

I kept giving everything a place because I hadn’t decided whether it still deserved one.

Every decision I hadn’t finished had quietly acquired real estate. A folder here, a database there, a recurring review, a dashboard panel, an AI thread. My system looked organized. What it was actually doing was faithfully preserving every choice I’d postponed.

Want to test one of yours? Pick a single system, the tracker or template or database you keep tending, and ask it four questions:

1. What were you supposed to help me do?

2. Do I use you, or mostly maintain you?

3. What decision is still sitting open inside you?

4. Can I finish that, stop it, park it on purpose, or pick the one place I’ll look for it from now on?

You don’t have to fix everything those questions turn up. One honest answer closed more for me than any reorganization ever did.

✳︎ ✳︎ ✳︎

What changed—and where you can start without copying me

Once I saw the real problem, the changes stopped feeling like a redesign. They felt like verdicts. The rituals I’d outgrown got retired for real. Not redesigned, not moved to a better app. Retired.

My weekly review stopped being another planning ritual. Now it looks at everything in motion and recommends what deserves movement, in what order, and why. The virtual board of advisors I’d built finally has a job there, but one rule holds it together: the advisors recommend. I decide.

Other decisions became just as plain. An AI conversation isn’t the record; when one ends in a decision, that decision goes where future-me will look. Each project gets one home, with duplicates removed or pointing there. And the dashboard stays local, read-only, and focused on pointing me back to the work instead of becoming more work.

Is it finished? No. I’m proving the weekly process by hand before I automate it. Otherwise, I’d just make the wrong thing faster.

You don’t need my arrangement. Pick one of these instead:

• Mark one stalled project as parked, on purpose, and take it out of your active view. Parked is a decision. Lingering isn’t.

• Pull one decision out of an old AI conversation and put it where future-you will actually look.

• Pick one project that lives in three places, choose the one place it really lives, and delete or redirect the rest.

One is enough. The point isn’t the system you’d be building. It’s the decision you’d finally be closing.

✳︎ ✳︎ ✳︎

Your study guide won’t look like mine

Back in school, a good study guide never learned anything for me. It didn’t replace the thinking, and it didn’t promise the kid next to me the same grade. It did one thing well: it made what mattered visible enough for me to make my own next move.

That’s all I’m really building now. A weekly advisory that distills what’s actually in motion without running my day. A Toolkit heading out to testers before every rough edge is sanded off. A dashboard that’s allowed to stay a simple place to get my bearings.

Yours won’t look like this, and it shouldn’t. You learn differently than I do. You decide differently. Your cheat sheet might be a single page you rewrite every Monday, a voice memo, a whiteboard, one honest question you ask before you open a single app. The shape doesn’t matter. What it helps you do does.

Because the clearest system may not be the one that remembers and displays everything. It may be the one that helps you use what you already know, and lets the rest stop asking for your attention.

Is there something in your setup that still exists because you’ve never made a firm decision about it?

✳︎ ✳︎ ✳︎

Want a first look at the Compass Check Toolkit?

I’m looking for a few beta testers. It’s still rough in places, and that’s on purpose: I’d rather have real eyes on it now than polish it in a vacuum. Message me if you’d like to be one of the first through it.

When a Popular Method Feels Like the Answer

A few years back, everyone was folding their socks into little rectangles. That was the Marie Kondo method, sparking joy everywhere. People were thanking their old t-shirts as they were putting them on the let-this-go pile. Her method became popular. It got shared, repackaged, and turned into some pretty cool businesses.

When I first heard about it, I thought, okay, I’ll try this. I’ll spark some joy. Then I went to my closet and realized this wasn’t going to work for me. I live minimally. It started after move one, and five moves later it’s still the same thing. I could see why it worked for people who consume a lot of stuff, but I wasn’t drowning in things. I didn’t need the method to work.

And that could have been the end of it. Not me, not my problem, move on. But there was still that little nag in my head of there’s something to this. I’m just not quite sure what.

So when the next thing showed up, I was ready for it.

This time it wasn’t sparking joy. It was Andrej Karpathy talking about building a wiki — a different kind of second brain you could actually navigate. Catalog everything. Link it together. Make it searchable months later.

That little nagging feeling from Marie Kondo came back immediately.

Maybe this was the thing I had been missing.

I didn’t have a physical clutter problem. I had a digital one. Or at least that’s what I thought.

For years I had accumulated conversations, notes, decisions, ideas, half-finished projects, articles, and experiments across more places than I cared to admit. Finding something wasn’t usually the problem. Finding my way back to why I made a decision six months earlier was.

A wiki sounded like exactly the thing I had been missing.

Finally someone had put words around a frustration I’d been carrying for years. It wasn’t just that my information lived in fifteen different places. It was that I never quite trusted I’d be able to find my way back to the thinking that got me there.

The pull felt real because it didn’t feel like another productivity system. It felt like an explanation for why I kept losing my way back to my own work.

So I opened ChatGPT and started building.

I thought I was about to build a better knowledge system.

The Moment I Realized I Was Building Someone Else’s System

At first, the conversation felt productive.

That was part of the problem.

ChatGPT was fast, competent, and full of structure. Every answer seemed to make perfect sense. Move this here. Rename that. Create an inbox. Separate active memory from the wiki. Decision records go here. Conversations go there.

I knew I had conversations everywhere—ChatGPT, Claude, Gemini, Notes—and somehow every version of the system still left me wondering where I was supposed to go when I needed to remember why I’d made a decision.

Nothing it suggested sounded ridiculous. Most of it sounded useful. And because it sounded useful, I kept going.

By the third conversation, we weren’t solving the problem anymore. We were reorganizing the solution we’d already reorganized once.

That was the part I had to pay attention to.

The answer wasn’t obviously wrong. It was aimed at a version of me that worked differently than I do.

But then a few steps in, something was wrong. Something wasn’t sitting right. I couldn’t quite name it yet. It was just that feeling that the thing we were building wasn’t going to work the way I needed it to.

I eventually went back and looked at those conversations. What struck me wasn’t that ChatGPT was wrong. It was how many times I told it I was still confused while we both kept building anyway.

✳︎ ✳︎ ✳︎

The Trap Isn’t AI. It’s Assuming the Misfit Is You.

Looking back at those conversations, I realized something uncomfortable. AI wasn’t wrong.

It wasn’t making things up. It wasn’t giving me nonsense. It wasn’t sending me in some obviously ridiculous direction I could dismiss and move on from.

That almost would have been easier. The harder part was that the answers sounded reasonable.

Your AI keeps guiding confidently somewhere that sounds useful. Your gut is quietly saying, this doesn’t quite fit. And you keep going anyway.

Because somewhere along the way, you decided the problem was probably you.

You’re thinking, this is a me problem. You haven’t found the right system yet. You haven’t explained it clearly enough. Everyone else seems to have it all figured out. When the AI guides you somewhere that feels a little off, you don’t stop. You push through because pushing back means trusting your own gut over the thing that sounds more certain than you feel.

And that feels riskier than maybe just being wrong.

That’s the trap.

Not bad information. Not a bad tool. Not even bad advice.

The quiet decision that the misfit is your fault.

And here’s where you let AI keep driving you to a destination that you already know is not really the place you want to be.

I almost did it during those three conversations with ChatGPT.

I had the feeling, and yet I kept building anyway, telling myself the confusion meant I wasn’t getting it.

I can see it now in the conversation itself. I wasn’t saying, this is working. I was saying some version of, I am no more clear now than before.

That’s the part I keep coming back to.

I wasn’t ignoring the feeling because I didn’t know it was there. I was ignoring it because the system sounded so reasonable.

And that might be the sneakiest part of working with AI.

It doesn’t have to be wrong to pull you away from yourself.

Sometimes it just has to sound more confident than you feel.

✳︎ ✳︎ ✳︎

What Changed When AI Started Asking Instead of Guiding

Eventually I hit a point where I couldn’t keep pretending the confusion was mine.

I finally typed:

“Stop acting like you know what is best and start getting clarification.”

It wasn’t written because I was angry with ChatGPT. It was written because I realized we were solving a problem I hadn’t actually agreed to solve.

That single sentence changed the conversation.

ChatGPT started asking instead of guiding:

• Where do the decision records need to live?

• What breaks if they move?

• What are you trying to find your way back to?

Those questions changed everything because they forced me to explain my thinking instead of evaluating its thinking.

At first I thought the breakthrough was better context. Looking back, I don’t think it was. The breakthrough was that ChatGPT stopped assuming it understood the problem and started verifying that it did.

Better context assumes AI already understands the problem and simply needs more information. Clarification questions test whether AI understands the problem at all.

I thought I was asking it to help me build a wiki. It turned out I needed help figuring out whether a wiki was even the right answer.

Those are completely different conversations.

✳︎ ✳︎ ✳︎

The Misfit Wasn’t the Problem. It Was the Data.

Looking back, that misfit turned out to be the most useful part of the entire experience.

I kept treating it like evidence that I hadn’t found the right system yet, when it was actually telling me I was solving the wrong problem. The wiki was never the answer I was missing. It was the catalyst that surfaced the question I should have been asking all along.

The questions started changing.

Instead of asking: “Where do I store this?”

I started asking: “If I need to remember why I made this decision six months from now, where would I naturally go looking for it?”

That was a completely different problem.

The question wasn’t where my information belonged. The question was how I found my way back to my own thinking. No popular method could answer that for me because it wasn’t really a knowledge management question.

It was a personal navigation question.

That’s what Marie Kondo started years earlier without me realizing it. Her method wasn’t wrong. Karpathy’s wasn’t wrong either. Neither one solved my problem because neither one was actually my problem to solve.

What they did do was expose the question underneath it.

That’s what I take away from this now.

When a new method shows up and everyone seems convinced it’s the answer, I don’t ask whether it’s good. I ask whether it’s solving the problem I actually have.

Those are two very different questions.

The misfit isn’t a verdict on you. It’s information.

When something feels off a few steps in, that feeling is data, not a sign you’re behind. The trick is catching it early, before you’ve built the whole thing, and saying it out loud instead of pushing through.

✳︎ ✳︎ ✳︎

How I Now Test Whether a Popular AI Method Fits Me

This experience changed something about the way I work with AI.

I don’t ask it to help me build the method anymore.

I ask it to help me decide whether the method deserves to be built in the first place.

When something catches my attention—a video, an article, a framework, a thread—I paste it into ChatGPT before I start reorganizing my world around it.

I want to try this, but I’m not sure it will work for me. Review the reference and find the gaps or inconsistencies with how I work. Ask me if you’re not sure or need to be more clear.

That last sentence is the part that matters most.

Ask me if you’re not sure.

It’s such a small change that it’s easy to overlook, but it completely changes the relationship. Instead of assuming it understands how I think, AI has permission—even the responsibility—to stop and verify it before it starts building a solution.

That one sentence also changes my responsibility.

Instead of evaluating the answer after it’s built, I’m asking AI to challenge its assumptions before we build anything together. If it can’t tell where the method and the way I actually work stop lining up, it has to ask instead of guessing.

That’s a very different conversation than simply asking it to make me a better system.

It’s a small shift, but it puts me back in the driver’s seat before I’ve invested three conversations building something that was never really mine to begin with.

✳︎ ✳︎ ✳︎

Before You Build Someone Else’s System

Looking back now, I don’t think Marie Kondo was ever wrong. Her method simply wasn’t built for me. Karpathy wasn’t wrong either. The wiki wasn’t a bad idea. It just wasn’t answering the question I actually needed to answer.

That’s the pattern I keep overlooking.

I treated each one like a test I was failing, when the real question was never whether the method was good. It was whether it fit the way I actually think and work.

I thought the lesson was going to be the system I eventually built.

It wasn’t.

The lesson was the moment I finally stopped building long enough to question whether I was building the right thing in the first place.

Now, whenever I feel that rush of “This is it. This is the missing piece,” I treat it as a signal to slow down instead of speed up.

Not because the method is probably wrong. Because I haven’t yet answered the more important question.

Is this actually solving the problem I have?

Because the question was never whether it works. It was whether it works for you, and you’re allowed to ask that question before committing yourself to someone else’s way of thinking.

That’s still one of my favorite ways to use AI—not to tell me what to do, but to help me figure out whether I’m solving the right problem before I spend weeks building the wrong solution. Sometimes the simplest questions are still the best ones.

“Is this actually for me?”

“Will this help me accomplish what I’m trying to do?”

“What assumptions am I making that I haven’t questioned yet?”

✳︎ ✳︎ ✳︎

If something feels off, don’t assume you’re the problem.
That’s exactly why I created the Something Feels Off waypoint—to help you slow down, sort through the friction, and figure out whether you need a different tool, a different method, or simply a different question.

I know that feeling. Not because of AI. Because of a book.

I’m a reader. It’s how I disconnect. No screen, no noise, just me and the page while ideas swirl in the back of my head. Most of what I read is fiction and biographies, because for me it’s about the story. And I have a rule: if a book doesn’t grab me, if it’s slow or it’s just not for me, I put it down and move on. No guilt. Life’s too short for a book that isn’t landing.

Then Lean In came out.

It wasn’t just a bestseller. It was a movement. Everyone was reading it, talking about it, building whole conversations around it. So I told myself, okay, I have to read this and see what all the fuss is about.

I had a hard time getting through it. Normally that’s my cue to stop. But this time I made myself finish, cover to cover, because of all the noise around it. Everyone seemed to be getting something out of it. What was I missing? Maybe if I just read the whole thing, I’d see what they saw.

I finished it. And the only thing I came away with was this: I don’t think this applies to me.

Which should have been fine. Except it didn’t feel fine. It turned into the question that actually stung. Why doesn’t this apply to me? What’s wrong with me that I can’t connect with the thing everybody else loves?

A Ladder With Every Rung Already Labeled

For a long time I didn’t have words for what bugged me about that book. I just knew it wasn’t for me, and I felt a little bad about that.

Then AI showed up, and the same feeling came back.

Look around at the AI conversation right now. Before most of us have figured out where it even fits in our work, there’s already a ladder built, every rung labeled. AI expert. Prompt engineer. AI strategist. Thought leader. The lists of top voices to follow. The “women in AI,” the “women in tech.”

I want to be careful here, because this part matters to me. Women in tech have been overlooked for a long, long time. The visibility is earned. The recognition is good. I’m not knocking anyone who’s climbing, and I’m not knocking the women who are finally getting seen. That’s not the problem.

The problem is what the ladder quietly says to everyone standing near it. That to matter, you have to climb. That if you’re not building the app, growing the audience, becoming the voice, then you’re behind. Or worse, that you’re holding yourself back out of fear.

✳︎ ✳︎ ✳︎

There’s Another Way to Lean In

Years later I finally understood what was bothering me about the message and the movement.

Lean In had one idea of success, and it was a narrow one. Climb. Get to the table. Get to the boardroom. And if you weren’t reaching for that, the book had an explanation ready: you simply hadn’t found your nerve yet.

But some of us weren’t afraid. We’d already looked at that version of success and decided it wasn’t the one we wanted. We’d chosen our own terms. Family, or a different pace, or work that fit the life instead of the other way around. The book couldn’t see that choice. It only saw a woman who hadn’t climbed yet. It invented a problem I didn’t have, and then handed me the guilt for not having it.

That’s the same move I feel in the AI conversation now.

There’s one shape of success on offer. Build the thing. Get loud. Become the authority. And if that’s not what you’re after, the explanation is already written for you: you must be scared of something, or behind, or playing it too safe.

I don’t buy it. And I don’t think you have to either.

Because there’s another way to lean in. You don’t have to lean in by going faster. You can lean in by going deeper.

When I say deeper, I don’t mean deeper into some specific app or model, learning every setting and every update. I mean deeper into the thing you’re actually trying to do. Pushing it to the edge of what you need it to do for you, and then stopping there. That’s enough.

And deeper is allowed to look small. It can be solving the single most time-consuming, annoying problem you have — the one that affects nobody but you, that nobody will ever clap for. It can be something you build to help you right now that’s useless in six months. That still counts. That’s still leaning in.

You took a chance on something small and unglamorous while everyone else chased big and shiny. Nobody hands out labels for that. Doesn’t make it matter less.

✳︎ ✳︎ ✳︎

What Lasts When Everything Keeps Moving

Here’s what I keep coming back to. Nobody has this figured out. The technology shifts every week, and outside the handful of people actually building it (it being the AI models), we’re all working with incomplete information. Me included.

That used to make me anxious. Now it does the opposite. If nobody has solid ground, then I don’t need to either. I just need to keep deciding what matters to me as things move.

I’ve lived through enough technology changes to trust that. In my twenties it went from analog to dial-up to modems. In my thirties, wifi and working online. Heading into my forties, online business became a reality for a lot of us. Every shift opened a door onto a dimly lit hallway. Nothing you could plan for, just feeling your way while you waited for more light to lead you.

Now there’s so much light you can’t figure out how to navigate. I’ve never felt anything move this fast. It’s exciting and it’s a little scary, both at the same time.

And what carried me through every one of those changes wasn’t being first to know, or having the most to say about it. It was being able to ask a decent question and decide what actually mattered to me.

That’s the skill that lasts. Not memorizing this month’s tools. Knowing how to decide.

So I have a few questions I ask myself now, almost without thinking, every time something new shows up:

Does this affect me?

Is it going to make my life better, or just different?

Does it touch my relationships, or the goals I’ve actually set?

That’s it. It sounds like me-me-me, and honestly, it is. I’m putting myself first. Not to get ahead of anyone, not to push some agenda. Just to know, for my own sake. Curiosity is allowed to be only curiosity.

And those questions don’t just sort AI tools. They sort anything. A medical decision. A legal one. Money. Anywhere the noise is loud and the stakes are yours, the work is the same. Figure out what’s actually for you, and let the rest go by.

✳︎ ✳︎ ✳︎

The Smallest Thing I Built

Where the idea came from.

My AI conversations kept going in circles. Sometimes it was because I’d picked up someone else’s prompt, and what it handed back didn’t match the way my brain actually processes information. Other times I gave it conflicting information, or sent it off on a tangent. But the bottom line is I was getting frustrated, because it didn’t seem like AI was the smartest book on the shelf.

How I applied it.

Back in late May, I wrote myself a short collaboration profile. A plain description of how I think and work:

• I figure things out by talking them through.

• I need one question or decision at a time.

• A wall of text or a big framework too early loses me.

• I’d rather react to something small than start from a blank page.

Here’s the actual ask I gave the AI to build it:

I want to create a short context doc for the way that I think, process, and work, because what’s happening is that with a lot of my asks, the waters are getting muddied. Thinking is different than processing, and a lot of the time the AIs get confused by that. I’m imagining something short and sweet that I can paste in or put in instructions, because a lot of my conversations are going sideways. When you and the other AIs formulate answers, it’s for John Q. Public, not someone who struggles a little bit like I do. Maybe we go through ten questions or so and then come up with a summary.

Now, whenever I’m working on something, whichever AI I’m in, I hand it that context first.

So that I’m not perpetually scrolling to find the advice or information I need. So that I get presented with options and time to think about them, instead of trying to decipher everything at once. Basically, I needed the AI to understand how I process information, so the output is specific to me.

The result.

And now it does the one thing I ask of it. The conversations stop circling. The AI meets me the way I process information instead of the way someone else does. It isn’t clever and it isn’t impressive. Nobody would look at it and jump up and down and go, why didn’t I think of this? I just use it every day. Instead of reiterating all of this in every conversation or every thinking session, I give it the parameters, and it knows how to have the discussion with me.

The tie forward.

And this is what I mean by building something small, just for you. If you’re sitting there feeling behind, wanting to build something big enough to count, this is the size of thing I mean. A plain document that makes one part of my day work the way I do. That’s a build. That’s what leaning in is to me.

✳︎ ✳︎ ✳︎

Back to the Book

So back to Lean In.

For years I carried that feeling from the finished book around as a small, quiet failure. Everyone else got something out of it, and I didn’t, and I let that mean something was off with me.

It wasn’t. The book had one idea of success, and it wasn’t mine. That’s all. The thing wasn’t wrong with me. The yardstick just wasn’t mine to measure against.

I think a lot of us are about to do the same thing with AI. Pick up the version of success that’s getting all the noise, hold our own quiet work up against it, and decide we’re behind. When the truth might be that we’re measuring with someone else’s ruler again.

This is the part I don’t want to lose. While everyone’s looking up at the loudest voices, the people actually changing things in our lives are often the quiet ones. The person who solved one real problem and mentioned how. The one who said here’s what I tried, and here’s what I found out, with nothing to sell you. We’ve been trained for a few years now to point our attention at the experts and the thought leaders. I wonder how many of the people making the biggest difference we’re walking right past.

You don’t have to become an AI anything. You don’t have to keep up, or get loud, or build something big enough to earn a label.

You just have to find the thing that’s actually yours, go a little deeper there, and share it if you feel like it. Not to get ahead of anyone. Just because it was yours, and it worked.

That’s leaning in too. It always was.

So here’s the one thing I’d leave you with, if you want it. That’s all I did when I wrote that little profile to help AI understand how I take in information and ask for what I need. It’s the kind of problem most people don’t even think to solve, and it’s quietly made the biggest difference in how I actually use AI.

So don’t go learn another tool. Pick the single most annoying, time-eating problem you have this week, the one nobody else would even notice, and see if AI can help you make one small dent in it. That’s the experiment. Not to have something to show for it. Just to find out.

✳︎ ✳︎ ✳︎

But sometimes the harder part isn’t picking the problem. It’s the nagging sense that you’re measuring against the wrong thing and can’t quite name what’s off. I built a short walk for exactly that. A few minutes, your own words, and you leave with a next move, not more noise.

Something feels off →

If you read DigiNav Compass Signal, you’ll stop second-guessing your AI and automation decisions because you’ll have a way to tell what actually fits your business from what’s just noise.

Subscribe now

The expertise is already yours. You just needed a framework to trust it again.

But some weeks an elaborate meal isn’t in the cards. It takes time and energy I don’t have.

A week or so ago I was at Costco and bought a bag of lemons. I had a picture in my head: a fresh lemon dessert, because doesn’t that sound like summer? Except who has time for that? Not me, not this week, not next.

So I had two options. Prep the lemons. Freeze the rind, freeze the juice, and have them ready for later. Or let them go soft in the bowl and do nothing with them at all.

Freezing it was.

And now every morning, before I head out to work out, I drop one lemon cube into my water. Lemon water. No extra work.

The dessert never happened. The thing I actually use every single day did.

✳︎ ✳︎ ✳︎

Hi, I'm Lee. I help solo business owners get clear on AI and automation decisions before they commit — no hype, no hustle, no borrowed playbooks.

New here? Find the room that fits → No Foundation Yet → Lost In Translation → Built For Someone Else → The Last Mile

Subscribe now

✳︎ ✳︎ ✳︎

The Noise

The same thing happens in business, and it’s costing people more than a bowl of lemons.

Everywhere you look right now, someone is telling you to build this or ship that. And here’s what most of us picture when we hear it: an app. A full automation. A system with seventeen moving parts. Something big, something technical, something that prints money while you sleep.

So you take that picture and you hold it up against your own week. And your week, the real one, doesn’t match. You ran maintenance for a client. You answered emails. You finally rewrote that one paragraph that had been bugging you. None of it looks like the picture.

And the quiet question starts: am I even up to this?

Here’s where it goes wrong. Somewhere along the way we started hearing “shipped” and translating it to “made money” or “went big.” We collapsed success into size. But those aren’t the same thing, and treating them like they are is how a perfectly good week starts to feel like a failure.

You’re two steps ahead of someone and two steps behind someone else. That’s true of everyone, all the time. The noise just makes it feel like you’re the only one standing still.

✳︎ ✳︎ ✳︎

What Shipping Actually Looks Like

So let me tell you what shipping actually looks like. Not the picture. The real thing.

Last week I ran maintenance on an Agency site. Updates, checks, the unglamorous keep-it-running work. That shipped.

I wrote and coded a waypoint for my own website, a small interactive process a reader can actually walk through. That shipped.

A proposal went out to a new client. Shipped.

I rewrote a web page, not because the old one was broken, but because a conversation with Claude got me clear on what I was actually trying to say. The new version went live. Shipped.

And this blog post — the one you’re reading — is a build-and-ship item. It started as a tangle of half-formed thoughts and it’s becoming a finished thing you can read. That counts too.

None of those made money the day they happened. Not one of them looks like an app or an automation. And every single one of them is a thing I built and put into the world.

That’s the part the noise erases. Shipped doesn’t mean big. It doesn’t mean monetized. It means finished enough to stand on its own, finished enough that someone could look at it. A maintenance run you can point to. A page that’s live. A post someone can read. The bar isn’t impressive. The bar is done and reviewable.

✳︎ ✳︎ ✳︎

Where AI Comes In

This is where AI actually earns its place, and it’s smaller than the hype wants it to be.

Most of the noise sells AI as the thing that builds the app, runs the automation, replaces the whole job. That’s the version that makes you feel behind, because it’s asking you to manage a new system on top of the business you’re already running alone.

That’s not how it’s been useful to me.

What it actually does is take something vague and heavy, the thing you’ve been circling for a week, and help you turn it into one small thing you can finish and look at. Not the blank page. Not “learn a whole new skillset before you’re allowed to start.” Just: here’s a rough draft to react to, here’s the page rewritten three ways, here’s the proposal you’ve been avoiding, now in front of you.

That rewritten web page I mentioned? I didn’t hand the writing to AI. I chatted with Claude to find where the ideas were thin or had gaps, and those responses let me rewrite the lines that were meh at best. Then I shipped the version that sounded like me.

The catch is real: this only works when the task is bounded and you know what good enough to ship looks like for you. AI will happily produce more of something generic, faster. It can’t decide what matters for your business or what sounds like you. That part is still yours. The win isn’t that AI did it for you. The win is that you no longer have to start from nothing to move one thing forward.

✳︎ ✳︎ ✳︎

The Tracker

Here’s what I did about all of this, because noticing it wasn’t enough.

I built myself a routine, a daily and weekly rhythm, to track what I ship. No matter how small. That was the whole rule: it counts, and it goes on the list.

Some days the entry is “ran maintenance for the Agency.” Some days it’s “wrote and coded a waypoint.” Some days it’s a proposal, or a paragraph I finally fixed, or a post like this one. The size of the entry isn’t the point. The entry is the point.

What that did, quietly, was take the argument out of my own head. When the noise shows up, and it always shows up, I’m not standing there trying to remember whether I actually do anything. I’ve got the list, and the list settles it.

It also changed what I reach for on a low-capacity day. Instead of staring at the elaborate version and doing nothing, I look for the smallest finished thing I can put on the list. Maintenance, not a rebuild. One paragraph, not the whole page.

Tracking it turned shipping from a thing I felt anxious about into a thing I could see myself doing. That’s the difference between a slogan you nod at and a practice you actually run.

✳︎ ✳︎ ✳︎

A Small Example, Since I’ve Been Talking About Small Things

Where the idea came from. There are weeks when I hit the same wall of what to share, because sometimes there’s no “big” experiment or project driving the topic. And a vague topic is no use to you. It isn’t anything usable, it’s just noise. The problem is never a lack of things accomplished in a week, or a shortage of ideas. It’s turning the vague or the mundane into usable raw material without creating cognitive overload.

How I applied it. On June 2nd, I built a dedicated GPT: the DigiNav Content Explorer. It doesn’t write articles and it doesn’t replace my process. It does one bounded thing: takes a vague topic and hands back a small, reviewable set.

• Question, Core Insight

• Audience Tension

• Potential Framework

• Audience Language

• Suggested Content Path.

The full research package is still there, but only if I ask. The default shifted from Question → Massive Research Package to Question → Insight → Decision → Optional Research.

So that I’m never stuck for too long on what to produce. Instead, the GPT lets me dig into the experiences I’ve already documented, or the questions that bother a lot of you. Sometimes it’s a memory jogger; other times it points me toward what to test next. This small build gives me a well-researched topic to follow in minutes, instead of wracking my brain over what’s next.

The result. It does what I built it for. The shorter output is easier to react to. I can look at it and move, instead of drowning in a research dump. It’s handled the topics I already had, the half-formed ones, and the flat-out “I’ve got nothing this week” mornings. Not a dramatic win. A small tool that does one job and gets used.

The tie forward. If you’re sitting there staring at the screen, wondering what you could share that might help someone else — this is the size of thing I mean. It’s not writing the article or inventing the conclusions for you.

It’s doing the heavy lifting on the research, and that’s often what sparks your next step.

✳︎ ✳︎ ✳︎

Back to the Lemons

So back to the lemons.

The dessert I imagined never got made. It was the big version, the one that needed time and energy I didn’t have that week, or the next. If “making something” only counted when it looked like that dessert, I’d have ended the week with a bowl of soft lemons and nothing to show.

Instead I’ve got a cube in my water every morning. Small. Finished. Doing its job with no extra work from me.

That’s the whole thing. The maintenance run, the waypoint, the proposal, the rewritten page, the GPT that hands me a topic in minutes, this post. None of them are the dessert. They’re all the cube. Small, finished, real, and already out in the world.

You’re shipping more than you think. You’re just measuring it against a picture that was never the point.

So here’s what I’d leave you with: stop checking your week against the big version. Think about the tiniest thing you put out this week. That’s a build.

Got something you need to think through?

If you’re stuck on which small thing to finish, or the whole thing just feels off in a way you can’t name yet, I built a short walk for exactly that moment. A few minutes, your own words, and you leave with a next move, not a pitch.

Something doesn’t feel right →

✳︎ ✳︎ ✳︎

Thinking Through Something?

If you’re sitting with a decision, workflow, tool stack, or AI process that technically works but doesn’t feel quite right, I offer limited advisory sessions for solo business owners trying to sort through the noise without adding more complexity.

Think It Through Together →

It’s not that I wanted to pick up her habits or her quirks. At the time I definitely didn’t want to raise five boys and three dogs. Okay, maybe the dogs. And the smoking and being put together 24/7 didn’t fit with the tomboy me.

But her confidence was something. And the way she handled the side-eye and the criticism — that was inspiring. I learned a lot from her.

Most likely it was due to the filters I had unknowingly developed. You know that gut reaction that tells you when something is off or not quite the right fit.

✳︎ ✳︎ ✳︎

You already know how to do this

Even all these years later, we still do this. Every day, with people and ideas and advice, we’re sorting what fits from what doesn’t.

But somewhere along the way it got harder to hear. The internet filled up with people who sound certain. The feeds, the newsletters, the courses, the YouTubers — everyone with a framework, everyone with a recommendation, everyone two steps ahead of where you are.

And it started to feel like a tool problem. Or an AI problem. Or a problem with not knowing enough yet.

It’s none of those. It’s a filtering problem.

✳︎ ✳︎ ✳︎

Why it’s so hard right now

Part of what makes this moment tricky is that almost no one is actually an expert yet.

AI is new enough that real expertise is rare. But you wouldn’t know it from the feeds. Everyone has positioned themselves as the person who figured it out. And when everyone sounds like an authority, the usual signals you’d use to sort them — credentials, confidence, follower counts — stop telling you much.

Then there’s the advice that’s just old. Techniques that made sense a year ago and don’t anymore. “Start every prompt by telling it to act as your role.” “Begin with the perfect prompt.” Some of it was never necessary, and as the tools evolve, more of it falls away. But it keeps getting repeated, because it once worked, and repetition sounds like authority too.

And the hardest one: plenty of brilliant people are sharing genuinely good work that still has nothing to do with your business. They’re not wrong. They’re just not building what you’re building.

That’s the part that catches you. Not the bad advice — you can spot that. It’s the good advice meant for someone else.

✳︎ ✳︎ ✳︎

Valuable is not the same as useful

So how do you sort the good advice meant for someone else from the good advice meant for you?

Not by deciding who’s worth following. By noticing the difference between valuable and useful.

Someone can be valuable. They expand how you think. They show you something you hadn’t considered. is this for me — much of what she shares isn’t built for my business, but the way she works changes how I see a situation.

Take her Open Door project, where a group each builds a piece of one collective space. Or the way she distills insights from others using AI and art. None of it is my use case. All of it shifts how I think about what’s possible.

That’s valuable. Valuable isn’t the same as useful.

Useful is when someone’s work consistently helps you make a better decision. This is also why the template habit can quietly work against you — borrowed structure that fit someone else’s process, not yours. Mia at is useful to me because she shares the thinking and the process, not just the finished build. I can take how she works through something and adapt it to my own business.

The value is in the method, not the output.

And then there’s the source you keep, even knowing most of it isn’t for you. Stu Jordan has been my AI go-to for years. He pushes the limits, further than I have any reason to go, because I’m not building what he’s building. I didn’t stop following him. I built a filter for him, a sense of what to actually act on and what to admire from a distance.

That filter is the whole thing. Not picking the right people. Knowing what to do with the ones you’ve got.

A few questions that do the work

Once you start filtering on purpose, a few questions do most of the work.

The first one I ask before building anything: why am I building this? Is it because it’s cool and I want to play with it? Because it actually moves something forward? Or because I want to learn the technique? All three are fine answers. The trouble is when you don’t know which one you’re in — that’s how “cool” quietly costs you a week.

The second is about who’s doing the recommending. When someone tells you a tool is the latest and greatest, it’s worth asking what’s in it for them. A lot of the time there’s an affiliate commission attached, or they’re building an offer around it. That doesn’t make them dishonest. It means the recommendation has a second job, and you should know that before you adopt it.

And the third is the one that stings a little. If you’re saving something for someday — a tool, a thread, a technique — but you’re not doing anything with it, is the save even worth it? By the time you come back, it may be outdated. The pile isn’t progress. It just feels like it. (I’ve written before about how saving outpaces doing — the gap between learning and acting.)

None of this means stop following people who are nothing like you. I do it constantly. It means take what they’re doing and bend it to your situation — your model, your clients, your use case. The signal that you’ve crossed a line: when their work makes you doubt yourself, or change course for no reason you can name. That’s not learning anymore. That’s drift.

✳︎ ✳︎ ✳︎

The prompt graveyard

I have a folder full of prompts I’ll probably never use. It’s the new version of the swipe file graveyard — collecting that feels like progress and isn’t.

I collected them the way everyone did. Prompt engineering was the thing, and we all became collectors — saving, screenshotting, copying every clever prompt we came across, sure that somewhere in the pile was the magic one that would unlock it all. I have prompts in there I never once tried. Others I saved are for a version of these tools that doesn’t really exist anymore. Old asks for an old moment.

That folder is what no filter looks like.

Not a character flaw — just what happens when you let everything in because it might be useful someday, and never run it through the gut check of is this mine.

So that’s the decision, and it’s a quieter one than it sounds. Keep following Alyssa, Mia, Stu, and the people who make you think. Keep learning from people who build nothing like you. But put it through the filter on the way in, not into a folder for later. Decide what’s yours and act on it now. Then put the rest aside — not because it isn’t good, but because good was never the question.

If you want to see your own filter at work, go look at the prompts you’ve saved:

✳︎ ✳︎ ✳︎

Back to the neighbor

I think about that neighbor sometimes. She never asked anyone’s permission to be exactly who she was, and she never tried to be anyone else either.

You have a version of her too. Someone you watched, somewhere, and knew without deciding it: that part is mine, that part isn’t. You didn’t study them. You didn’t save them to a folder. You just took what fit, and moved on from the rest.

That instinct is still in there. It’s the same one you’re trying to use on every expert, every tool, every thread you’ve saved for someday. You’re not missing a skill. You’re trying to remember one you’ve had since you were a kid, watching someone three houses down be entirely themselves.

✳︎ ✳︎ ✳︎

Thinking Through Something?

If you’re sitting with a decision, workflow, tool stack, or AI process that technically works but doesn’t feel quite right, I offer limited advisory sessions for solo business owners trying to sort through the noise without adding more complexity.

Think It Through Together →

I could’ve gone to the store and bought prints. Easy. But I wanted something that was mine. Something I’d put my own blood, sweat, and tears into.

But painting is hard work. It takes time, and it takes talent. So I found my way around it: paint-by-number, with my own flair on top.

At first I felt weird hanging them up. It’s a paint-by-number. What am I doing putting this on my office wall?

Then people saw them, and nobody knew. Everybody assumed I’d painted them from scratch, which was the craziest thing.

There I was, feeling like a painter imposter for taking the easy way, adding my little bit of flair, passing it off as my own art. When that’s exactly what it was. I was just in my own head about it. Hand me and somebody else the same numbered canvas and the two would still come out different. Nobody looking at it ever saw the numbers underneath. They just saw the picture, and either it caught their eye or it didn’t.

The same thing happened when I decided to move my personal website off WordPress. The site is a hub. It points to my two businesses (DigiNav Compass and WP Site Success) and to where I write on Substack. That’s the whole job — one page, three doors.

✳︎ ✳︎ ✳︎

The decision should have taken a week.

It took six months.

Not because the technology was difficult.

Not because the messaging wasn’t ready.

Not because the design was unfinished.

Those were the reasons I kept giving myself, but none of them held up under scrutiny.

I kept telling myself it was the messaging. Then it was the imagery. Then the design. Then whether I should host it on Vercel or Cloudflare. Every time I solved one question, another appeared.

And every time I got close to launching, I’d find another reason to wait.

Something felt wrong.

At least that’s how I described it at the time.

What I know now is that the feeling wasn’t warning me about the project. It was pointing at something I didn’t want to look at.

For eighteen years, I’d been the WordPress person.

I’ve built sites. Maintained sites. Recommended WordPress to clients. Built an entire business around it.

Now I was putting my own website on something that wasn’t WordPress.

That’s what I was resisting. Not the migration. Not GitHub. Not Cloudflare. Not the code.

The story I had about who I was.

The realization didn’t hit until I was decommissioning the WordPress version of the site. Removing the maintenance plugin. Taking it off the server. Closing the chapter.

That’s when I finally understood why it had dragged on for six months.

I wasn’t evaluating a website decision.

I was grieving an identity.

And once I saw that, I knew what to do.

✳︎ ✳︎ ✳︎

If a client came to me stuck like this, I’d start ruling things out. Is it the messaging? The design? The functionality? The audience? What’s actually broken?

So I turned the questions back on myself.

What evidence am I waiting for that I can only get after launch? If someone else showed me this site today, would I tell them to launch it?

What specifically feels unfinished? If I had to launch this Friday, what would I remove from the checklist?

What am I having trouble letting go of? Which parts of the old site represent years of work rather than current value?

Am I stuck because the site isn’t ready, or because launching forces me to admit who I am now?

If you’re stalling on something that should be easy, sit with the last one.

That last question was the one.

The site was ready. I wasn’t. Or more accurately, my feelings hadn’t caught up to a decision I’d already made.

✳︎ ✳︎ ✳︎

Here’s the part nobody puts in the articles or the YouTube videos.

Embrace AI. Vibe-code this. Build an app for that. That’s the easy part. The tools are easier than they’ve ever been. The friction isn’t there anymore. You don’t need permission, you don’t need a developer, you don’t need much of a budget.

What nobody talks about is the part that comes before you touch any of it. The part where you have to come to terms with not doing things the way you’ve always done them. Where the method that built your business isn’t the method that fits this job. Where using the easier tool feels like betraying the harder one you’ve spent years on.

That’s the work. Not the tool. The unlearning of who you were before the tool existed.

And the kicker, for me, was that I’d been telling people this for two years. AI should meet you where you’re at. Match the solution to the need. It’s the line I open conversations with. It’s the line I close them with.

I just hadn’t applied it to myself.

The principle didn’t fail. I almost did, because I was so busy walking the walk for everyone else that I forgot the walk includes me.

✳︎ ✳︎ ✳︎

It made its debut on time. Deadline was set and nothing was going to stop the train from leaving the station.

The site is what I’d pictured. Static HTML, no heavy maintenance, hosted on a server I use regularly, pointing to the places I can be found. A routing page doing the job of a routing page, with my voice and my vibe.

What I didn’t expect was the self-doubt rearing its ugly head for that long. I overlooked the imposter, burn-the-bridges feeling I was sitting with for all those months. If I’d asked myself the hard questions in month one or two, month six would never have rolled around.

And if I’d been my own client, I’d have fired myself for the flippy, floppy indecision. The good news is I gave myself a second chance. The better news is I’ve now got a better diagnostic for the put-this-off, pit-of-my-stomach feeling.

✳︎ ✳︎ ✳︎

If this tension feels familiar, you may want to keep exploring from here.

Related Theme

The Last Mile: The final stretch — where the work is done and the only thing left in the way is you.

Thinking Through Something?

If you’re sitting with a decision, workflow, tool stack, or AI process that technically works but doesn’t feel quite right, I offer limited advisory sessions for solo business owners trying to sort through the noise without adding more complexity.

Think It Through Together →

I’ve never tried any of them.

Who wants to take an ice-cold shower first thing in the morning? Who wants to wake up at 4 or 5 if they don’t have to? But the articles got me thinking about what my morning actually looks like.

I don’t get up with an alarm. Usually 6:15 or 6:30. I work my way into the day. Personal hygiene. First cup of coffee. Brain puzzles — Wordle, the crossword, a few others. Second cup of coffee. Some casual reading — sometimes a Substack, sometimes a newsletter, sometimes a book, depending on the mood. Then the gym.

No email. No client problems. No digging into the day’s productivity.

That part is on purpose. If my day starts chaotic, my whole day is whittled away.

So I protect the quiet before the noise gets in.

✳︎ ✳︎ ✳︎

A method to my madness

That same instinct is what made me notice something was off with the rest of my week.

I’ve been building a lot lately. Daily process. Weekly process. Decisions database. AI is the collaborator on most of it — I lean on Claude the way I’d lean on a smart co-worker who has all my context. The systems are supposed to do for the rest of the day what the morning routine does for the start of it: keep the noise out, surface what matters, let me make the call.

But every single thing I’d built lately had a friction point.

Not one of them. Every one.

Projects were starting and never finishing — sitting on a shelf, not abandoned, just stuck. The kind of stuck where you keep meaning to go back to it and don’t.

My first read was the easy one: I’ve outgrown this.

That was wrong.

But I didn’t know it was wrong yet. What I had was two different systems sending the same kind of static, and the only way I could think to start was to type the static into a chat window and see what came back.

Going outside Claude

My first instinct was to burn one of them down. The experiments database had already been shelved. I hadn’t touched it in weeks. I was ready to add the rest to the pile.

I didn’t.

Instead, I went outside my normal partner. Claude has all my context, which is great most days and a problem on this one. Sometimes you need someone without the history. So I opened ChatGPT, pasted in the instructions I’d written for each of my Claude workflows — the actual prompts that drive what Claude does when I run my daily process or my weekly process — and then pasted in documentation of how my brain operates.

Not personality. Mechanics. The way I produce lists badly and evaluate lists fast. The way storage works fine but retrieval breaks. The way I think differently when I’m given options to react to versus a blank page to fill.

ChatGPT told me the instructions were fighting my flow.

Not the instructions themselves. The instructions, for the way my brain works.

That was the moment the processes changed. And ChatGPT was the partner I worked through the refactors with — the workflow instructions and brain documentation already in the context, different shape of input, different shape of output.

✳︎ ✳︎ ✳︎

The prompts that gave me the diagnosis

Each refactor started with a sentence I typed into ChatGPT. Not a question — a noticing. The kind of sentence you write when you can feel something is off but haven’t decided what to do about it yet.

Daily process.

I started with a simple ask. A question on what I’d noticed, without being specific:

The signal: running the daily rhythm felt like pulling teeth.

The daily was supposed to look at where work actually lives — calendar, client logs, project list, decisions — pull from all of them, and funnel the surfacing into one place where I could make a call. Instead, it was inventing tasks. Asking me what I wanted to do today and then double-checking I had the time and energy to do it.

What’s the point of running this if I’m going to do all the work anyway?

That’s the moment a system stops being collaborative and starts being theater.

The core tension underneath: the system was becoming work instead of supporting the work.

Weekly process.

With the daily process sorted, it was time to tackle the next:

The signal: I was trying to step up a level — from weekly to monthly to quarterly — and the review wasn’t a place to think. It was a place where decisions had already been made.

The weekly was supposed to look at what may actually move the needle, name the growth items, and break them into bite-size chunks for the week ahead. What it was doing was producing a tidy summary that closed loops without leaving room to see anything. By the time I read it, the thinking was over.

The core tension underneath: the review was deciding for me instead of helping me reflect.

A status report, not decision support.

✳︎ ✳︎ ✳︎

Two different leaks. And underneath, the same misread.

Outgrown is not the same as never fit

I’d assumed friction in a system meant I’d outgrown it. The system used to fit, and now it didn’t, so the system had to go.

What was actually happening: I’d never fit any of those systems to me in the first place. I’d built them off what other people were doing, off templates that work for other people’s brains, off whatever the latest article or video or thread suggested. Some of it I was conscious of. Most of it I wasn’t.

The friction wasn’t because I’d outgrown it. It was because I’d never adjusted it to how I think.

The thing is, I’d been watching this same pattern play out somewhere else.

When you feed a model junk context, it gives you junk output. The model isn’t broken. The retrieval isn’t broken. The shape of what you fed it is the problem. The whole conversation drifts because the input was the wrong shape from the start.

I was doing that to myself.

My experiments database was structured like a software change log — every decision, every step, every iteration, every detail. Useful if you’re publishing it. Terrible if you’re trying to reorient yourself a week later. Too much information, too much noise, too much minutia. Junk context.

The system was working exactly as designed. The design was the problem.

The shape of someone else’s template

Stay with me on this part for a minute.

I didn’t sit down one morning and decide to build my daily process on someone else’s template. Nobody does. The templates get absorbed the way background noise gets absorbed. A productivity article here. A YouTube workflow there. A thread that walks through how somebody runs their week. You’re not copying any of it. You’re just exposed to it. Then you sit down to build your own version, and the shape that comes out of you is already half-shaped by what you’ve seen.

The daily was supposed to be mine. But the idea of a daily — what one should do, what it should produce, what it should feel like to run — that came from somewhere else. Same with the weekly. Same with the experiments database. I’d built each of them in response to something I’d read or watched, and I’d done it without noticing the inheritance.

That’s the part I missed the first time.

Friction in a system you built yourself reads like I’ve outgrown this because you assume you knew what you were building. But you were never the only architect. The template was in the room with you. Some of what you built was yours. Some of it was the shape of how everyone else does it. And when the friction shows up, it’s almost always at the seam between the two.

The morning routine works because I built it from nothing. No one handed me a 6:15-coffee-puzzles-gym template. I watched what worked, kept what kept, dropped what didn’t. The reason it doesn’t have friction isn’t that I’m gifted at mornings. It’s that there was no inherited shape to fight.

The workflows had inherited shape everywhere. That’s why every one of them leaked.

Each system has a job to protect

Once I could see the inherited shape for what it was, the refactor got obvious. The morning routine protects quiet before noise gets in. The workflows weren’t protecting anything yet — they were just running.

So I asked the question differently for each one. What is this supposed to protect?

Daily — protects energy and capacity.

Look at the calendar. The client logs. The ticket system. The projects I’m working on in both businesses. And the decisions I’m tracking. Pull from all of those, funnel it into one place, and let me make the call on what’s actually getting done today. The run should help me decide, not hand me a list of made-up tasks. If it feels like one more chore in the day, I’ve built it backwards.

Weekly — protects the growth thread.

What actually moved this week? What were the growth items? What are the bite-size next steps for next week? The weekly is supposed to help me see what happened so I can decide what’s next. If it’s handing me the answers before I’ve had a chance to look, it’s the wrong tool for the job.

The refactors weren’t replacements. Neither one got burned down. I just rebuilt them to work the way I actually work.

✳︎ ✳︎ ✳︎

The surprise

Then the experiments database — the one I’d shelved, the one I was ready to burn — turned out to be the rescue.

I started looking at it through the same question: what is this supposed to protect?

And the answer wasn’t every detail of every experiment.

It was the decision I made, and the context I’d need to make a similar one next time.

It wasn’t an experiments database. It was a decisions database with the wrong sign on the door. What I actually needed was something I could come back to, get my bearings, pick up the context without having to remember every detail. And on top of that, something I could pull from when I’m sharing what worked or what didn’t — in an article, a video, or a conversation with someone trying to sort through the same thing.

Experiments → Decision Records.

The data I’d already captured wasn’t wrong. The frame around it was. That covered the records of what I’d already decided. The companion to it came from a different place.

Over the past few weeks, I’d been writing down the questions I kept asking myself when I got stuck — the ones I couldn’t quite shake. I had ten of them scattered across notebooks, Notion pages, and the back of receipts. How do I know what’s truly next? How do I know what work fits the energy I have today? How do I make decisions when my brain reacts faster than it retrieves? Those became the Decision Playground. A small Notion database where I can pull up the right question for the kind of decision I’m trying to make — what direction to head, what signals I’m noticing, what my capacity actually is, what I’m watching, what I’m reflecting on.

When I get stuck on what’s next, I have somewhere to start. When I open a conversation with AI for help thinking it through, I have the right context already shaped.

The system I was about to delete was the one I needed most.

I just hadn’t called it the right thing yet.

Replace, refactor, or refine

The hardest part wasn’t realizing something was off. It was figuring out what to do about it.

I had three options on the table for every system I touched. So do you, the next time something feels off.

Replace — burn it down, build something new in its place. This is the option I almost took. It’s also the option most people take, because friction feels like failure, and the response to failure is start over.

Refactor — keep the bones, rebuild the way it works underneath. The system was doing the right work, but the shape was wrong. You’re not throwing the work out. You’re asking the system to do its job differently.

Refine — small adjustments inside the existing design. Tighten a prompt. Move a field. Reword a question. The lens is right. The system just needs a tune.

Most of what I read online treats these as the same thing. They’re not. Picking the wrong one costs you weeks.

Here’s how I tell them apart now. You can use the same test.

Replace when the purpose is wrong. Not the structure, not the output — the underlying job the system was built to do. If you’ve outgrown the need the system was built for, replace. Note: this is actually rare. Most things you think you’ve outgrown, you haven’t.

Refactor when the purpose is right but the frame is wrong. The system was built for the right reason, but the way it’s structured is off. The experiments database belongs here — the purpose (capture what I’m doing) was right. The frame (changelog) was wrong. Renaming and restructuring revealed what was already there. No new work to invent.

Refine when the purpose and frame are right but the execution is rough. A prompt that produces almost-right output. A field that asks the wrong question. A step that’s in the wrong order. These are small. You’d know in a session whether the refine worked.

The trap is treating everything like a refine when it needs a refactor — because refine feels efficient and refactor feels expensive. Just one more tweak is the most expensive sentence in a stuck workflow. If you’ve made three refines in a row and the friction is still there, you don’t have an execution problem. You have a frame problem.

The reverse trap is treating everything like a replace because the friction makes you angry. Replace is the option that costs you the most work and is right the least often. If you’re reaching for it, slow down. Ask first if the purpose is genuinely wrong, or if you’re just frustrated with the shape.

The systems I refactored this week looked, at first, like they needed to be replaced. None of them did. Yours probably won’t either.

✳︎ ✳︎ ✳︎

What I’m watching for

The refactored systems are running now. The honest test isn’t whether they made sense on paper — it’s whether they hold under regular use, or whether new friction shows up that signals the next misfit. I’m not assuming this is solved.

Friction in a system is a signal worth paying attention to — not a verdict, but information.

✳︎ ✳︎ ✳︎

If you’re sitting on a shelf

If you have a system, a workflow, a tool, a habit, a planner, a recurring meeting — something you’ve quietly stopped using but haven’t replaced — that shelf is data. Whether it’s a Notion database or a paper notebook, the question is the same.

Try this. The same thing I did for the systems above.

1. Write the noticing as a sentence.

Not a fix. Not a question for AI yet. The shape of something is off and I haven’t decided what to do about it. Mine looked like “this feels like a chore to run” and “there’s so much reference to everything else.” The sentence is the signal that something’s there worth looking at.

2. Ask what it’s actually doing for you — not what the label says it’s for.

The honest answer sometimes means rebuilding the whole thing. My experiments database wasn’t just mis-labeled — it was holding the wrong shape of data for the job it was actually doing. Once I saw it was a decisions database, the structure had to change too: different fields, different way of capturing the entry, different way of retrieving it later.

Other times the label is fine and the system itself needs to do its job differently. Daily rhythm was a fine name, but the system was generating tasks when it was supposed to be helping me decide. Weekly review was a fine name, but the system was wrapping things up when it was supposed to be opening them up so I could see them.

Asking what something’s actually for doesn’t always end in new vocabulary. Sometimes it ends in a full rebuild.

3. Before you rebuild, ask what it’s supposed to protect.

This is the question I missed the first time around. A system that doesn’t protect anything is just busywork. A system that protects the wrong thing leaks where it matters. The morning protects quiet before noise. The daily protects energy and capacity. The weekly protects the growth thread.

If you don’t know what your system is supposed to protect, you’ll rebuild the same misfit with new paint.

✳︎ ✳︎ ✳︎

Whether it’s a morning or a workflow, it comes back to the same thing.

Protect what’s important.

The noise will find its way in either way.

✳︎ ✳︎ ✳︎

If this tension feels familiar, you may want to keep exploring from here.

Related Theme

Built for Someone Else: Building workflows that fit your actual business and energy instead of someone else’s blueprint.

One thing to sit with: What system have you quietly stopped using — and is it actually outgrown, or was it never built for the way you work?

✳︎ ✳︎ ✳︎

Thinking Through Something?

If you’re sitting with a decision, workflow, tool stack, or AI process that technically works but doesn’t feel quite right, I offer limited advisory sessions for solo business owners trying to sort through the noise without adding more complexity.

Think It Through Together →