Is Consulting a Dying Industry With AI?

Ask Google whether consulting is a dying industry and you will get an AI Overview that says, roughly, “not dying, but changing.” That is true and also useless. It does not tell you whether to hire someone, or what to hire them for.

Here is the sharper version.

What AI actually killed

It killed the pyramid.

The traditional model ran on volume. A firm put a lot of junior people on a project, billed their hours, and those hours went into gathering data, building models, and formatting slides. That work was real, it was slow, and it was expensive. It is now close to free.

I can produce a competitive landscape, a first-pass financial model, and a clean deck in an afternoon. So can you. So can the client. That is the whole disruption, and pretending otherwise is how firms end up defending a price no one believes anymore.

The second thing it killed is generic advice. If the recommendation you are about to deliver is something a client could have gotten by typing the question into ChatGPT, you do not have an engagement. You have a subscription they are overpaying for.

What AI did not touch

Three things, and they are the three things that were always the actual job.

Judgment in a messy room. Every organization I have worked in has a version of the same problem: two leaders who disagree, a budget that does not cover both, and a team waiting to see who blinks. No model resolves that. Someone has to sit in the room, read what is not being said, and make a call that people will actually follow.

Accountability. Executives do not only hire outside help for information. They hire it to share risk. When a hard decision goes sideways, there is a difference between “the model suggested it” and “we brought in someone who has done this before and they own it with us.” That is not a service AI can provide, because AI cannot be held to anything.

Execution that survives contact with real people. This is the one I care most about, because it is where most AI projects die. The pilot works. The demo is great. Then it hits a team that has a workflow, a manager who did not ask for this, and a quarter to hit. Getting from “it works” to “we use it every day” is a change management problem wearing a technology costume.

The part nobody says out loud

Consulting is not shrinking. It is separating.

On one side, the work that AI does faster and cheaper, which is heading toward zero. On the other, the work that requires a person who has run something, not just advised on it. Those two things used to be bundled and billed together. They are coming apart, and the middle is where the pain is.

If you are a buyer, this is good news. You can now buy the second thing without paying for the first.

What this means if you are hiring

The old vetting questions do not work anymore. “How many people will be on the team” used to signal capacity. Now it mostly signals overhead.

Ask these instead.

What will be different in ninety days, and who owns it after you leave? If the answer is a document, keep looking. If the answer names a workflow, a person on their team, and a way to tell whether it is working, that is a real engagement.

Have you run this, or only recommended it? There is a large gap between someone who has advised on adoption and someone who has sat through the meeting where a team says they are not going to use it. I have been on the wrong side of that meeting. It changes what you build.

What are you going to do first, and why that? This is the one that separates people fast. Most failed AI adoption is not a tooling problem. It is an order problem. Teams buy the model before they have cleaned the process it is supposed to run on, then conclude AI does not work for them. It is the whole reason I built The Sequence Model, and it is the first thing I look at on any engagement.

If you want the longer version of that vetting conversation, I wrote it out in How Do I Choose an AI Consultant?. And if you are asking this question because you are wondering whether to become one, the market side is covered in Is There a Demand for AI Consultants?.

So, dying?

No. But the safe version is over.

A consultant whose value was knowing things is in trouble, because knowing things is now a commodity. A consultant whose value is judgment, ownership, and staying through implementation has more demand than ever, because the number of organizations trying to adopt AI badly has never been higher.

That is not a comfortable answer for the industry. It is a very good one for anyone buying.

If you are trying to figure out where AI actually fits in your organization and in what order, that is the work. You can see how I structure it on the offerings page, or just book time and we will look at your actual situation.

Book a 30-Min Strategy Session

Consulting is not shrinking. It is separating.

FAQ