AI didn’t break consulting. It just made bad consulting visible.

I use AI every day in my work. Research, analysis, first drafts, and sense-checking my own thinking. It’s saved me hours I used to spend on things that, if I’m honest, weren’t where my value was anyway.

So, when I say what I’m about to say, I want to be clear: I’m not sceptical of AI. I’m sceptical of what some people are using it to avoid.

The narrative that needs questioning

The story doing the rounds is that AI threatens the consulting industry. That the profession is under pressure, that fees can’t be justified the way they used to be, and that clients will increasingly do it themselves with the right tools.

There’s something in that. But it’s missing the more uncomfortable observation.

AI didn’t threaten consulting. It exposed the parts of consulting that were never worth the fee in the first place.

What AI genuinely does well

Give it credit where it’s due.

AI synthesises information faster than any consultant alive. It pulls research, organises frameworks, spots patterns across large volumes of data, and produces a first draft of almost anything in minutes. It doesn’t get tired, doesn’t have an off-day, and doesn’t bill you while it’s thinking.

I use it for exactly these things. And I’d encourage any consultant who doesn’t to start, because the clients who figure out that you’re not using it will start wondering what they’re paying for.

But here’s the thing about all of that: it’s the cheap part.

The hiding place

For years, a significant portion of consulting work was expensive information retrieval dressed up as insight.

Research compiled from sources the client could have found. Frameworks pulled from a textbook, relabelled, and presented on slides with the client’s logo on top. Recommendations were written in the passive voice so that nobody was clearly accountable for anything.

Clients paid for it because it looked thorough. Because the deck was 40 slides. Because the consultant had a name-brand firm behind them and spoke confidently about industry benchmarks.

AI does all of that faster, cheaper, and without the day rate. And if that’s what a consultant was selling, they’re in genuine trouble.

They should be.

What 30 years in the room actually gives you

I’ve been doing this for 30 years across industries and company sizes. And the thing I’m most certain AI can’t replicate isn’t empathy or human connection (though both matter). It’s something more specific.

It’s knowing what the data isn’t telling you.

AI works with what’s in front of it. It synthesises what exists. What it can’t do is walk into a business, sit in a leadership meeting, and notice that the operations director hasn’t made eye contact with the CEO once in 45 minutes. It can’t read the silence when you ask about staff turnover. It can’t feel the gap between what the numbers say and what the room is telling you.

That gap is usually where the real problem lives.

A few years ago, I was brought in to help a business that believed it had a sales problem. The data supported that framing. Conversion rates were down, the pipeline was thin, and the sales director was under pressure.

Within a week, it was clear the sales team wasn’t the issue. The product had drifted from what the market wanted, and nobody at the board level was willing to say it out loud. The sales team were trying to sell something that had stopped being the right answer.

No amount of AI-generated research would have surfaced that. It took conversations, observation, and the willingness to say something the board didn’t particularly want to hear.

The uncomfortable truth

Which brings me to the second thing AI can’t do.

AI is trained to be helpful. Specifically, it’s trained to give you what you asked for, in a form you’ll find useful, without unnecessary friction. That’s a feature when you want a research summary. It’s a problem when your client needs to hear that their strategy is sound, but their leadership team is the issue.

The willingness to say the uncomfortable thing isn’t a technical capability. It’s a character question. And many consultants fail for the same reason AI does: because the path of least resistance is to tell people what they want to hear.

I’ve lost work for saying things clients didn’t want to hear. I’ve also had those same clients come back 2 years later because nobody else would. That’s not something you train a model to do.

What this means if you’re hiring a consultant

If you’re evaluating a consultant or wondering whether you’re getting value from one you already work with, here’s a useful question to sit with.

Strip out everything they’ve produced that AI could have generated. The research, the frameworks, the slide deck. What’s left?

If the answer is a clear diagnosis of the problem you couldn’t see yourself, a conversation that challenged how you were thinking about it, and a recommendation with a specific owner and a measurable outcome, you’re in good hands.

If the answer is a well-formatted document and a confident delivery style, you might want to ask harder questions.

AI raised the floor for mediocre consultants. Faster outputs, better-looking decks, and more comprehensive research. But it did nothing for the ceiling. The things that separate good consulting from expensive research retrieval were always human. They just weren’t always visible.

They are now.

Have a brilliant week!

Dave Rogers, The Business Explorer

If this struck a chord, my Substack goes further on the same questions. Want help applying any of it to your business? Book a call.

Need a speaker who’ll bring a fresh perspective? Reach me at info@fuelledfitandfiredup.com.