TL;DR
Navi helps me keep track of emails and projects while I work on other things. I still need to make the ask clear, review the work, and decide what I want to keep doing myself. We're both learning. The question is whether that help saves me time overall, or whether I'm spending more time managing it.
The turkey arrived at dessert time
Thanksgiving is right around the corner, and in our family that means one big coordinated dinner.
Whoever is hosting cooks the turkey, the potatoes, and the stuffing. Everybody else brings the side dishes. Then it all comes together.
My daughter hosts now. And instead of just roasting a turkey, they like to also deep fry the turkey.
So now there's even more chaos to the process.
The first year we did it, one part of the meal was done before the other part. We ate. Some people were ready for dessert. And here comes the deep-fried turkey, because the timing was all off.
So the next year, we tried to get a better grasp on the timeline of all the food being cooked.
Coordinating these big get-togethers is something I'm really good at. We have a large family, and any family meal or get-together is a production. Normally I'm the one doing the majority of the cooking, or the prep for whatever everybody contributed.
The problem with the deep-fried turkey was that I had prepared the turkey and then gave it to my son-in-law to cook. Not that he couldn't cook it. But I was handing it off to someone with no idea of:
how long it would take to cook
how long it needed to set up
what we were going to do to carve it
After the first year or two of trial runs, it works really well now. The turkey gets prepared two different ways, plus everything else, and it all comes together into one cohesive meal.
The meal itself never changed. There would be a protein, the sides, the desserts. What changed was a second cooking method. We just needed to figure out how to work it into the plan.
That's the same thing I noticed when working with AI agents. You hand something off, it does the work, and it comes back so it can be coordinated. The first couple of times, it might be a little clunky and need to be refined, because it's a new part of the process.
✳︎ ✳︎ ✳︎
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.
Today, I’ll be showing you what an agent actually takes off my plate, what it still takes to manage it, and how the handoff is changing.
✳︎ ✳︎ ✳︎
What has actually left my plate?
Everyone is telling you to build agents and let AI do the work for you. Building bots or agents is totally useful, and I love it.
Navi is my dot, one of the Dots agents in ChatGPT. I tell it what to do, and it gets the job done. It's also proactive. It tells me what I need to tackle, or which emails came in that need my attention. It's like having a personal assistant.
The email reminders are a good example:
What Navi does: tells me which emails need my attention, so I don't have to keep checking everything.
What I keep: I still decide what to say, and I have control of the sending of the messages.
Navi also keeps track of my projects, and what is falling away, sitting too long, or taking unneeded attention.
That help shows up in my week. With AI and my agents doing the minutiae, and a plan in place with solid guidelines, there's extra time for other things. The Monday transition that used to be scattered and stressful is now just calm and productive.
But handing work off is still work. There's the conversation to set priorities, the check back when it's done, and the redirect when something isn't what I asked for. None of that means the agent made a mistake. It's part of working with it.
So the question for me, and maybe for you if you've handed work to an agent or a scheduled task:
Is the agent or automation saving you time, or are you just managing it more?
✳︎ ✳︎ ✳︎
What to think about: The work you hand over and the work you keep
Working with Navi usually goes like this. We have a conversation, we prioritize, I hand off the things that can happen, and then I go about my business and check back when it's done.
That's the real benefit. While it's working on some things, I can work on others.
But what gets handed over has to be clear. This morning, Navi got several tasks from me at once:
check email, prioritize what needs responses, and see if we can draft them or take care of them
tell me what needs my attention
the article I need to write today, which starts with reviewing what I did last week to see what the topic can be
reviving my YouTube channel, starting with a series of "in plain English" videos
Navi came back and assumed the article and the videos were going to sit together. That wasn't the ask. So I had to respond, say that wasn't right, and redirect it.
And that was just as much me as it was Navi.
Sometimes three things get thrown together in one ask. That's a problem for the agent, and it's a problem for me:
My side: I need to get clearer on separating ideas.
Its side: it needs to get clearer on the ask, and say something like, "These are three ideas rolled into one. Do you mean this?"
So I told it: if I'm not clear on my ask, you have to call me out on that.
It's like when you have an employee or a team member. You have to get used to each other's flow, the way you ask for something, and when you have to clarify or not.
So when you look at whether an agent is saving you time, look at both sides. There's the work it takes off your plate. And there's the work you keep: being clear about what you're handing over, and making room for it to ask when it isn't sure.
✳︎ ✳︎ ✳︎
What can go wrong: Acting before the request is clear
When the ask isn't clear, the agent spends time doing something I didn't ask for. Then I have to spend time reviewing and redirecting it. That's why I want it to clarify the ask before it starts, if something isn't clear.
But there's another side to this.
I made the ask: respond to this email with what I wanted to say. It created the response and sent the message before I had a chance to review it.
Luckily, I told it what I wanted to respond, so it wasn't a big deal.
But think about what could have happened:
It might not have used my own words.
It might not have understood the response I was trying to give.
And I didn't tell it not to send, or to draft only.
That could have worked against me.
Client emails are the part I don't want to hand over completely. I don't want to remove myself. Even though the agent has my voice, it might not describe something the way that I would. The context might be off. Or I might have a more personal relationship with the client, and the email would seem a little too not me.
The agent didn't break a rule here. I never gave it one. It acted on an assumption, and so did I.
So we have to make sure that there are no unchecked assumptions, on either side.
✳︎ ✳︎ ✳︎
What comes next: Agree on the work before it starts
Now I'm a little more clear at the start. This is what we're going to work on today. And if there's anything I'm not clear on, or you don't understand, let's clarify it first.
Dots also have a voice option. It's kind of like a phone call, where you call in and talk to your dot.
That call is where things get kicked off now, because it's more like having a conversation. We talk about what we're going to handle today, and it always starts the same way:
Let's prioritize today.
Tell me the things that need to be done.
Then tell me where you can step in.
That last part matters. It helps me get more used to the capabilities, instead of making the assumptions.
None of this is finished. It's still learning, and so am I. I don't expect it to get everything right immediately, but I do expect the corrections to lead to less work for me over time.
If you're working with an agent or a scheduled task, the start of the work is a good place to look. Agree on what it's handling, what you're keeping, and what needs to be clarified before it begins.
Is this making the work easier?
I'm not changing course. Navi stays.
What changes is how the work gets handed over:
Clearer asks: separating my ideas, instead of throwing three things together.
Clarify first: if something isn't clear, we sort it out before the work starts.
What I keep: I still decide what to say to clients, and I control the sending.
The bigger question is still open: is the agent learning and making things easier over time, or am I still doing the same amount of work managing it?
There's no number for that yet. Tracking time would help, both on the things taking my attention and on how much I'm saving or the AI is using. That's how to see if this is really something that's making a difference.
Have you used an agent like this yet? Maybe one of the Dots in ChatGPT, Meta's Muse, another personal autonomous agent, or a scheduled task that runs on its own.
Tell me in the comments: is it saving you time, or are you managing it more?