On this page
The one AI habit to model so your team actually copies it
Pick one visible habit, do it this week, and narrate it. That is the whole plan.
What you'll have when you're done
One real AI habit you personally run every week, made visible to at least one member of your team. Not a mandate. Not a training. A behavior your team can see and copy.
The mandate nobody follows
You have probably already given the AI speech. Maybe you sent the email. Maybe you announced it at the all-hands. Here is how that usually lands: a handful of people who were already using it get a little more permission, and everyone else waits to see whether you actually mean it.
The mandate without the behavior is everywhere right now. Leaders announce AI adoption, sign the vendor contracts, and then keep doing their own work exactly the same way. The team clocks it. Not consciously, but it registers: this is not real yet.
Change-management research is consistent on this one point: leader modeling is the primary driver of behavior adoption. The common finding across successful AI transformations is that the CEO changed first, visibly, and the team followed. The mandate accelerates nothing if the person who issued it is not doing the thing.
This is actually good news. You do not need a rollout plan yet. You do not need a champion network or a training calendar. You need ONE habit you actually do this week, in a way your team can see.
What you need first
- A paid Claude or ChatGPT account you already have.
- One task from your week that takes 15 minutes or more: a meeting, a report you have to read, a document you have to write.
- Enough context to narrate what you did. Two sentences is plenty.
That is it. No new software. No setup time.
Step by step
Why modeling beats mandating
Teams copy what leaders DO, not what they announce.
When a CEO sends a memo saying "we're an AI-first company now," the team files it under Things Leadership Says. When the same CEO opens a meeting with "I ran the agenda through AI last night and it pushed back on my framing of the second agenda item, so I rewrote it," the team registers something different. The change is real. The cost of trying it just dropped.
A visible leader habit does two things at once. It signals the change is serious enough for the most senior person in the room to actually do it. And it lowers the social cost for everyone else. Nobody wants to be the first one to show they tried something new and got it wrong. When the CEO already did it and named what got corrected, the door opens.
The most reliable adoption lever you have is not a mandate. It is your behavior.
The one habit to pick
Pick one. Run it this week. These are the three I would choose between:
1. Weekly meeting prep
Before a meeting you already have scheduled this week, paste the agenda or email thread into a plain chat window and ask it to pressure-test your assumptions. Ask what the other side is likely to push for. Ask what questions you should be ready to answer.
Then open the meeting with: "I ran this through AI last night. Here is what it pushed back on."
That line does a lot of work. Meeting prep in chat walks through the exact prompts and the five-minute protocol.
2. Pre-read summarization
You have a long report, a deck, or a thread sitting in your inbox. Before the meeting where it comes up, paste it into chat and ask: "What are the three decisions I need out of this?"
Then share the summary alongside the original when you distribute materials. Drop the prompt in the notes. Show the work. Summarizing documents and threads is the step-by-step for this one.
3. One weekly draft
Your Friday update, your board memo, a team note. Write a first draft yourself, then hand it to AI to sharpen: tighter openings, cleaner structure, places where you buried the lead.
Show the before and after to one direct report. The diff is the whole point. Drafting emails and updates in chat covers the approach.
Pick the one you will actually do this week, not the most impressive one. The habit that runs on Monday is worth more than the sophisticated habit that runs once in a quarter. Frequency and visibility beat sophistication every time.
How to make it visible
Doing the thing is the first half. The second half is narrating it. Four moves:
Name it in the room. "I asked AI to pressure-test this before I came in." One sentence. Not a speech. Just narrate what you did. The team hears that you are doing this as normal work, which means they can too.
Share the artifact, not just the story. Drop the prompt and the output into a shared Slack channel, a shared doc, or the meeting notes. Not a summary of what happened. The actual thing. When people see a real before-and-after, or the exact prompt you used, it is suddenly trivially easy to try the same thing.
Show what you changed. This is the one most people skip, and it is the most important. The edit is the credibility. When you show that you looked at what the AI gave you, threw out two paragraphs, rewrote the third, and kept the fourth, you are demonstrating judgment. You are not outsourcing the decision. You are using a tool and making the call. That is the thing your team is watching for.
Invite one person to try the same thing this week. Not a directive to the whole team. A one-to-one invitation, by name, to one person whose work you know would benefit. "Hey, I used this before the board update this week. Think it would work on your monthly wrap? Want to try it?" A direct invitation spreads faster than a mandate because the person receiving it feels seen, not managed.
How you'll know it's working
Someone on your team narrates a habit back to you. "I tried that meeting prep thing you showed us." That is the tell. They are not following a policy. They copied a behavior.
The questions change. Instead of "should we be using AI," your team starts asking "how did you prompt it for that?" That is the shift from permission-seeking to adoption.
You start keeping the habit. The genuine test is whether you still do it four weeks from now. If the habit was real to begin with (low-stakes, high-frequency, something you actually needed), it survives. If you picked it to perform, it dies. The team will notice either way.
When it breaks
Mandating before you model. If you ask your team to use AI while keeping your own process unchanged, the team sees the gap. Leadership says one thing and does another. That gap is exactly what creates shadow compliance, where people route around the tool and tell you what you want to hear. You cannot close that gap with a better mandate.
Performing instead of using. Pick a real task you actually need done, not a task you chose because it looks good. Teams are good at detecting theater. A CEO who ran AI on a task they clearly did not care about sends a worse signal than a CEO who did nothing. Pick something that matters to you, something you would have done anyway, and just add the AI step.
Starting with something high-stakes or embarrassing. Low-stakes, high-frequency is the zone. Start with prep for a routine meeting, not the board deck. Start with your Friday team update, not the investor memo. The goal is to build a reflex, not to make a statement. The high-stakes version is easier once you have run the reflex fifty times.
Making it a big announcement. The habit should look like normal work you happen to narrate. A grand announcement puts it on a pedestal, which means any stumble reads as a failure. A casual mention, "I prepped with AI for this one," keeps the bar right: this is just how work gets done.
Level up
This piece is the starting point: one habit, one week, made visible to your team.
Once you have run the habit a dozen times and you are seeing your team copy it, the next step is a structured rollout. That is a different gear: use cases mapped to roles, champions assigned, rituals adapted, adoption measured.
How to get your team to actually use AI is the intermediate piece for when you are ready for that layer.
If you want a full 30-day plan, rolling out AI in 30 days has the sequence.
Start with the one habit. Run it this week. Narrate it once. Tell me in thirty days what changed. I'd love to hear about it.
Andrew
Get three workflows like this every Thursday
The Thursday 3 is a free weekly email. Three workflows that put you in the top 1% of CEOs. 90-second read. Every card links back to a step-by-step guide like this one.
Get the newsletter →The architecture behind this workflow.
Two operator manuals for the same job, run two ways: OpenCLAW for the always-on harness, Claude Code for the focused-work CLI. Pick one, or get the bundle for $149.
Browse the books · $99 eachWant one workflow like this taken apart end-to-end every week? The Tuesday Pro Deep Dive · $39/mo.