Hi Reader,

The senior leaders I work with all hit the same wall with AI.

They give it a half-formed instruction. They expect it to back-fill the context, infer what they really need, ask one or two sharp questions, and come back with something they'd actually use.

Because that's how it works with a smart team. You say "can you take a pass at this?" and a good engineer or PM or consultant fills in the gaps. They know your business. They know your call. They've watched you make decisions for two years. They write back with the version you would have written, if you'd had the bandwidth.

AI doesn't do this. Not yet. So senior leaders give it the kind of direction they used to give a senior IC, and they get back 20 pages of bullet points and frameworks. Technically fine. Completely useless.

The skill that made you senior is the exact skill that makes AI feel broken.

What I'm Noticing

The people getting the most out of AI aren't asking better questions. They've made one structural reframe: they stopped expecting AI to read between the lines, and started spelling out the things they used to delegate to context.

This is uncomfortable at first. It feels remedial. You're explaining the format. You're capping the length. You're telling it what NOT to do.

Then it works. And the work compounds.

Your Next Move

Three translations from how you used to lead people to how you have to lead AI:

1. What you used to say: "Can you take a pass at this?" → What works with AI: "Write me one paragraph. Then stop. I'll respond with a paragraph of my own. We'll go back and forth until we have it." Forces turn-taking. Prevents the firehose response.
2. What you used to say: "Make it tight." → What works with AI: "Don't go past 200 words. Ask me before adding a new section." A team member knows what "tight" means. AI needs a number and a checkpoint.
3. What you used to say: "Why did you do it that way?" (after the fact, to someone you trusted) → What works with AI: "Where did you go off? What should I have told you up front to keep you grounded?" The answer is your prompt template for next time. You're not retraining the model. You're training yourself to lead it.

Do this next week: Pick something AI has frustrated you on lately. Before you re-prompt, ask it the third question. The answer will tell you more about how to lead AI than any course will.

Life Beyond the Screen

There's a parallel in how I work with my kids.

When my older one is melting down, I used to hit her with three questions at once. "What happened? Are you okay? Do you want a snack?" That was me trying to solve faster, not actually meet her where she was.

What works is one short question. One beat of silence. One response from her. Then the next thing.

Senior leaders are good at this with humans, by the way. You do it intuitively with a junior who's overwhelmed. The thing that gets in the way with AI is that it doesn't LOOK like it needs that pacing. It looks like it can take everything at once. So you give it everything at once. And you get a mess.

A question for this week: Where in your work are you treating AI like a smart team member who'll figure it out, when you should be treating it like a brilliant intern who needs scaffolding?

Work with me directly. If you want a thinking partner on your product strategy, AI rollout, or building your fractional practice, here's how I work with people: https://gamma.app/docs/Chill-Labs-Coaching-Information-ifjuyb0hjgw9xtb

Let's Chillaborate,

Dina

Founder, Chill Labs

PS: New episode of Automate Yourself drops June 29 with Danielle LaFleur. Follow the show on Spotify so you don't miss it: https://open.spotify.com/show/3BMSWEdcGbIfXXQM6YoVdQ 💙