The Four Habits That Make You AI-First
- Amy Westlake
- Jul 21
- 4 min read
Situation
Ten people on a call, most cameras off, one hour on the calendar. A team I'd never worked with had asked me to come talk about AI. Word had gotten around from my AI use case Slack channel, and by then a handful of teams I had no real connection to were reaching out the same way. In a prep call beforehand, I asked the manager what she wanted out of the session. She didn't have a list of tools in mind. What she told me was simpler: her team didn't know what they needed, didn't know what questions to even ask, and didn't know where to start.
Nobody in that room was resistant. They just didn't have a place to start. So instead of the tools overview, I built the session around four specific habits.
The AI Move
Habit 1
Block ten minutes to play. No task attached, nothing to show for it afterward. Just follow whatever you're curious about that day. I'd written about this before, and I repeated it here in case you didn't see the previous post about this or you needed a little reminder.
Habit 2
Start with AI before you start the document. Instead of opening a blank document and staring at it, open an AI chat conversation instead, out loud if you can. Use the microphone and just talk through what you're thinking. The structure shows up on its own once you stop trying to type it into existence.
Habit 3
Use something you already like. This one has nothing to do with AI specifically. It's just how I learn anything new, and I do it every time. When I was figuring out Claude's Cowork mode, I didn't test it on a work deliverable. I planned a trip with it instead. I already know what good trip planning looks like: what a real itinerary needs, what a lazy one skips. So I could tell in ten minutes more about how the tool worked rather than focusing on the content.
Habit 4
Ask AI to teach you something, instead of asking it to produce something. I tried this with a relative of mine who's in his eighties. He was curious about cryptocurrency, not because he wanted to invest, just because he wanted to understand how it actually worked. I asked AI to write a short, plain-language explainer aimed at someone with zero background in it, and gave that to him. He didn't stop there. He kept the thread open and started asking his own follow-up questions. He'd never used AI for anything before that.
The Shift
Living these four myself didn't produce a single moment I can point to. What it did was keep confirming the same thing, over and over, session after session. Ten minutes most days kept turning up something I hadn't tried before. Nobody handed me a reveal. I just kept finding out there was more here than I'd gotten to yet, which is a strange thing to keep learning and still be a little surprised by.
I love learning, so that part came easy for me. Watching it land on that call, and later watching it land with my relative, is what made me want four specific habits instead of one vague encouragement to use AI more. A room that doesn't know what to ask needs something smaller to hold onto than "use AI more."
The Pattern
None of these habits ask for more time or a better tool. They ask you to move AI earlier in your process, before the task feels important enough that testing something new on it seems risky. Habits one and two get you there: something low-stakes to reach for at all, and a habit of reaching for it before the blank page instead of after.
Habits three and four are about what happens once you're already reaching for it. Pick something familiar enough to judge fast, a trip itinerary you already know the shape of, and you have the chance to better understand the tool. Or hand someone a plain answer to something they were too unsure to even ask about, and watch whether they keep going on their own. My relative didn't need a lesson plan. He needed one explanation and permission to keep asking.
The Implication
Pick one habit. If the ten-minute habit is already running, try opening your next real task with a conversation instead of a blank document. If nothing's started yet, take something you'd do anyway, plan a trip, pick up a hobby, and run it through AI before you hand it anything that matters. Or find someone who's curious about something they've never had explained well, and see what happens when you hand them one good answer instead of a syllabus.
What I'm Testing Next
I'm going to start asking every team afterward which of the four actually stuck, if any. Four might be the right number. It's also possible only one of them does the real work and the others are just noise. I won't know until I ask.