Your Client Already Talked to AI — That's Good News

Your Client Already Talked to AI — That’s Good News

I’ve noticed a shift. Every week, the topics my clients bring to our sessions have already been through multiple rounds with AI. They’re good at using it — they have their own frameworks, custom skills, and they’ve gone deep on the questions that matter most. Then they come to me.

I’m usually hearing about the problem for the first time. No prep. Straight into it.

This isn’t just a consulting thing. Your colleagues, your boss — they’ve probably talked to AI before bringing a problem to you. And it’s happening more and more.

Sounds like bad news for the human expert. After six months, I’ve found the opposite.

Calibration Is the Real Value

There’s a counterintuitive finding from Dr. Google research: informed patients made doctors spend MORE time with them, not less. Doctors had to correct the wrong conclusions patients picked up online.

But here’s the thing — doctors say this “correction” work is actually the most valuable part of the visit. Explaining basics? Anyone can do that. Finding which specific conclusion doesn’t hold in YOUR situation? That’s expert work.

Clients don’t bring correct answers from AI. They bring a set of judgments that need calibration. And calibration updates beliefs. That’s where the value lives.

Risk-Taking Beats the Median

Being unprepared can be an advantage — but only because you bring something else into the room. You know where it might be possible to take a risk. That’s a uniquely human judgment.

Roger Martin’s insight: AI produces median-level answers. It infers patterns from massive data and finds the mean. It doesn’t look for exceptions. Doesn’t even try.

Your client talked to AI for five rounds and got a logically sound median plan. Every step makes sense.

But to beat the median, you need to find the spot where taking a risk pays off. That’s exactly what AI won’t do.

Non-Consensus Frameworks Are the Only Moat

My experience: even though every client has already discussed the problem with AI, I can still help them see it differently through a unique framework. Not because I know more — AI knows far more than I do. It’s because I see the problem from a different angle.

Socratic questioning. Jobs’ design principles. Munger’s mental models. Buffett’s value analysis skill. All great. All on GitHub.

Easy skill replication has made these thinking methods average for a certain crowd. Roger Martin puts it bluntly: if your framework is average, you’re in AI’s kill zone. The only moat? A perspective that makes your client say “I’ve never thought about it that way.”

Your unique angle, layered on top of proven frameworks, produces completely different results.

The Consensus Concerns Are… Consensus

AI sycophancy. AI accountability gaps. Collapsing information asymmetry. 70% of consulting work being automatable. You’ve probably read all of this. It’s all true — but it’s all consensus. Your smart clients can learn this from AI itself.

The question worth your time isn’t “will AI replace me?” It’s “after my client talks to AI for five rounds, what’s the thing that’s still missing?”

Back to That Moment

Since my clients started using AI deeply, we can focus on what actually matters: underlying beliefs, system goals, paradigm understanding, structural design.

AI didn’t make consulting shallower. It made it deeper.