Reading Time: 6 minutes

Hey Prompt Lover,

Harvard and BCG sat 244 professional consultants down and watched every single thing they did with AI.

Every prompt. Every response. Every decision.

4,975 interactions in total.

And what they found should make every person who uses AI for work stop and think about what they're actually doing.

Because here's the thing.

Every single person in that study had the same AI tool.

Same task.

Same data.

Same instructions.

But what they did with it? Completely different.

And those differences had nothing to do with intelligence.

Nothing to do with experience.

Nothing to do with how good they were at their jobs.

It had everything to do with one thing.

How they chose to work with AI.

Some people treated AI like a teammate. They pushed back when it got things wrong. They added data to test its thinking. They disagreed out loud. They questioned it. They refined it. They were in the conversation the whole time.

Others used it more like a reference book. They'd ask it specific questions when they needed to look something up, then go do the actual work themselves.

And then there was a third group.

They copied and pasted everything into one prompt and let AI do all of it. Problem framing. Analysis. Recommendation. The full thing. They did a quick check at the end to make sure it didn't say anything crazy, and they submitted it.

Fast. Polished. Done.

And here's what made this interesting.

The third group thought they were winning. They finished faster. The output looked good. Clean formatting, confident language, professional structure.

But they got the answer wrong more often than everyone else.

And more importantly — and this is the part that should really make you think — they didn't learn anything.

Not about their work. Not about AI.

Nothing.

The researchers gave these three groups names.

The ones who worked hand in hand with AI the whole time — constantly in dialogue, constantly steering, constantly questioning — they called them Cyborgs.

The ones who used AI selectively, for specific things, while doing most of the thinking themselves — they called them Centaurs.

The ones who handed everything over and stepped back — they called them Self-Automators.

These aren't insults. They're patterns.

And what makes this research genuinely useful is that each pattern produces a different outcome. Not just in the quality of work you produce today. But in who you're becoming as a professional.

Cyborgs are getting better at working with AI.

Centaurs are getting better at their actual job.

Self-Automators are getting better at neither.

Here's what I need you to sit with before tomorrow's issue.

Ask yourself honestly: when you use AI at work, which one are you?

Not which one sounds best. Not which one you think you should be. Which one are you actually doing right now, on a regular Tuesday afternoon, when you have a deadline and you just need the thing done?

Because most people, if they're really honest, are somewhere between Centaur and Self-Automator. And they don't always know the difference.

Tomorrow I'm going to break down exactly what each type looks like in a real work session — with specific examples so you can see yourself in one of them.

And the day after that I'm going to tell you what to do about it.

One more thing before I let you go.

The Claude guides are starting April 2nd.

Full walkthroughs. Cowork setup. Claude Code. The phone to laptop bridge. Real setup instructions, not surface level stuff.

If you've been waiting for that — it's coming.

See you tomorrow.

Attio is the AI CRM with conversational AI built directly into your workspace. Every morning, Ask Attio handles my prep:

All in seconds. No searching, no switching tabs, no manual updates.

Ready to scale faster?

Start now →

— Prompt Guy

P.S. Reply and tell me which one you think you are. Cyborg, Centaur, or Self-Automator. I'm genuinely curious.