I get asked this at least once a week now. Usually it arrives as a message at about eleven at night, from someone with twenty-five years in their field, and the message is a link to an article and one word: “Thoughts?”
I know that message. It’s from someone brilliant who has started to feel the ground move under them.
The honest answer sits between the two lazy ones, the reassurance and the apocalypse. The market is changing shape, and the new shape decides which end of it you should be standing at.
Will AI replace consultants
No. It will replace consulting work that’s average, repeatable and answer-shaped, which is most of what the middle of the market sells. Judgement formed over decades, access to rooms, and accountability for a decision stay human.
The middle is collapsing. The top and the bottom are both growing.
What artificial intelligence actually replaces
It replaces the retrieval of known answers. Frameworks. Benchmarks. First drafts of a strategy document. Market summaries, the competitor round-up, the twelve-slide deck that reorganises what the client already told you. All of that used to be billable because it was slow. It’s no longer slow.
It replaces the junior layer of the pyramid, which is the commercial engine of the traditional firm. The model where a partner sells and a bench delivers depends on the bench costing less than the work. When a tool can do part of the bench’s job, the arithmetic of the whole firm changes.
And it replaces the consultant whose real product was information asymmetry. If your value was that you knew what good looked like in an industry the client couldn’t see into, that moat has been drained dry.
The World Economic Forum’s Future of Jobs Report 2025 puts numbers on the direction of travel. Forty percent of employers anticipate reducing their workforce where artificial intelligence can automate tasks, and workers can expect 39 percent of their existing skill sets to be transformed or outdated between 2025 and 2030.
That second figure is the one people misread. Skills go stale. Judgement doesn’t.
Why the numbers don’t say what people think
The US Bureau of Labor Statistics projects employment of management analysts, the occupational category that holds most consultants, to grow 10 percent between 2025 and 2035, against 3 percent for all occupations. There were 1,077,100 of those jobs in 2025. The category is growing faster than the labour market around it.
Put the two findings side by side. Demand for advice is rising. The cost of producing average advice is falling towards zero. When those two things happen at once, the market doesn’t disappear. It stretches. Volume floods to the bottom, where the number approaches nothing. Value concentrates at the top, where it can’t be substituted. Everything in between gets squeezed out.
The field is fine. A particular position inside it is dying, and it happens to be the position most consultants were taught to build.
I wrote the book about exactly this. The middle is collapsing, and it isn’t coming back. Iconic Influence is the full case, and it came out this month.
How to tell if you’re standing in the middle
The word middle is doing a lot of work in this argument, so I’ll make it concrete. Your revenue doesn’t decide it, and neither does your seniority. The middle is a description of how substitutable your offer is.
Four questions settle it.
Could a competent competitor deliver something your client would consider close enough?
Is your number arrived at by comparison, against an hourly rate, a competitor quote or an internal budget line?
Is most of your deliverable a document, a framework or an analysis?
Would your client struggle to name what, specifically, they’d lose by hiring someone else?
Every yes moves you towards the middle. Four yeses and you’re likely standing in the part of the market that’s being removed, however good the work is and however senior you are. Quality doesn’t protect a substitutable position. It only makes the substitution slower.
Are you with me? Because none of those four questions mentions artificial intelligence. The middle was always structurally weak. What’s changed is that the weakness now has a fast, cheap, tireless competitor sitting inside it.
Why the top is the one place it cannot reach
Four things resist substitution, and none of them is knowledge.
Accountability. A board can ask a tool for a recommendation. It can’t ask the tool to be answerable for it in eighteen months. Somebody has to own the call, and ownership is a human transaction with a reputation behind it.
Access. A large part of what a serious client buys is the room you can put them in. Introductions, alliances, the phone call that gets returned because of who’s making it. No model has a network. It has a corpus.
Discernment under ambiguity. Machine output is confident and average by construction. The hardest calls in a business are the ones where the data is thin, the incentives conflict and the answer is a judgement about people. That’s where decades of pattern recognition earn their number, and where a plausible answer is more dangerous than none.
Presence. At a certain level, the client isn’t buying a document. They’re buying somebody sitting across from them who has done this before and isn’t frightened. In a world of artificial everything, authentic human mastery becomes the ultimate luxury.
What this means if you’re a genuine master
You’re probably being told to compete with people half as good, on their terms, using their tools. That advice was written for a volume business, and for a volume business it’s good advice. You don’t run one.
If you have decades of real depth, the response to this market is to move up, deliberately, and let the middle go. Fewer clients, far bigger engagements, and no less wealth. Work that’s commissioned rather than bought off a shelf. Your positioning has to change before your number does, because a premium number on an undifferentiated position is just an expensive version of the same problem.
Lindsay Mustain is my favourite picture of the move. She was “successful” inside the mass market of human resources, by every external measure, and she knew she was the best in the world at what she did. The gap between that knowing and the work she was selling was unbearable. I know that gap. I lived in it for years in my first business, successful on paper and quietly suffocating, so when someone describes it to me I don’t need the long version. She moved up instead of speeding up. Her words now: “I am charging hundreds of thousands in each transaction thanks to you, Kathryn. You are some crazy magician.” I love Lindsay. She’s a big part of why I do this.
Same Lindsay. Same mastery. Different altitude.
It also means being honest about which half of your work is genuinely irreplaceable. Most experts are carrying a service line that a tool now does better, and subsidising it out of pride. Let it go. It was never the reason anyone hired you.
Separate what only you can do
Start by separating your work into what a machine can produce and what only you can. Be ruthless. The first list is longer than you want it to be.
Do this on paper, engagement by engagement. Take the last five pieces of work you delivered and break each into its components: the research, the synthesis, the recommendations, the document, that sort of thing, and then the conversations, the decisions you carried, the relationships you brought. Then mark each one honestly. Most experts find that between half and two-thirds of what they delivered is now producible at speed by a tool, and that the value the client actually remembers sits entirely in the remaining third.
That third is your business. The rest was scaffolding you were charging for because it took time.
Price the second list as though the first no longer exists, because commercially it doesn’t. This is where most people stall, and the stall is a positioning problem. If your market can’t tell the difference between your judgement and a competent output, the fault sits in how you’re positioned, and positioning is the most buildable thing you own.
Build the proof in public. The argument you hold is now a bigger differentiator than the credential you hold, because credentials are checkable and arguments can’t be copied.
Change the shape of the offer before you change the number. People run this in reverse, and reversing it is what produces the experience of raising your rate and watching your pipeline empty. If the offer still looks like a document delivered over eight weeks, a higher number is a worse deal for the same thing. If the offer contains access, accountability and a decision the client can’t make without you, the number stops being the subject of the conversation.
And accept the smaller client list. Six or eight serious engagements a year, run properly, is a different business from thirty run at a sprint, and it’s the one that survives a market where average output costs nothing. The strange part, the part almost no one believes until it happens, is that the work gets better as the list gets shorter. And you get your weekends back. I’m not being cute. You do.
The machine wants your throughput. Your clients want your judgement. Sell the second one.
If you want to see what your judgement is actually worth once the scaffolding comes off, the Iconic Pricing Calculator is the start. Five minutes. Free. Discover what you could be charging.
Kathryn x
FAQ
Will AI get rid of consulting?
No. It removes the parts of consulting that were paid for because they were slow: retrieval, synthesis and first-draft analysis. Demand for advice itself is rising, and the US Bureau of Labor Statistics projects the management analyst category to grow 10 percent between 2025 and 2035. What disappears is the middle position, and the field carries on without it.
Is AI a threat to consultants?
It’s a threat to a specific business model, the pyramid where juniors produce billable output that a tool now produces faster. Accountability, access, judgement under ambiguity and presence sit beyond its reach, and those four things are what serious clients are actually buying at the top of a market.
Is consulting a dying field?
The data says otherwise. The category held 1,077,100 jobs in 2025 and is projected to grow faster than the average occupation over the following decade. What’s dying is the undifferentiated middle, where the work is comparable, the number is negotiable and the value was information the client couldn’t get elsewhere.
