The Empty Register: Why Nobody Is Incorporating a Management Consultancy Anymore

The most telling data point in professional services this quarter did not come from a McKinsey white paper or a Davos panel. It came from the driest possible source: the UK Companies House register. AIBD analysis of SIC code 70.22 filings shows zero new management consultancy incorporations in Q3 2026, a 100% decline against the prior period. When the formation rate of an entire industry category hits zero, something structural has happened. Aspiring founders have looked at the market, run their numbers, and declined to play.

That verdict deserves attention.

The Pyramid Was Always the Product

To understand why, it helps to remember what the consulting business model actually was. It was never, strictly speaking, about insight. Insight was the advertising copy. The product was the pyramid: an elegant labour arbitrage in which junior analysts gathered data, mid-level consultants built the narrative, and senior partners sold the relationship. The pyramid scaled beautifully. It generated enormous margins. And it depended, crucially, on information asymmetry: the client did not have the analytical horsepower the firm had.

AI has not merely nibbled at that model. It has walked directly into the engine room.

As one recent analysis put it plainly, AI now threatens multiple layers of that structure simultaneously, with tasks that once required teams of junior consultants working long hours now completable dramatically faster using generative AI tools. That is not a productivity improvement for the existing firms. It is the elimination of the entry-level economic rationale on which the entire apprenticeship model was built.

The Split Nobody Wanted to Say Out Loud

The consulting market is dividing, and the division is not subtle. According to the latest Management Consulted industry report, the market is bifurcating between scaled ecosystem integrators and specialised execution firms. Large firms are deepening partnerships with hyperscalers and enterprise software providers, while specialist firms are gaining share through narrow expertise. The generalist middle, the comfortable two-partner boutique advising on "digital transformation" without doing any of the transforming, is being hollowed out.

This is Clayton Christensen's disruption playbook, running at double speed. The low end of the market (junior research, synthesis, deck production) has been captured by AI tools. The high end (relationship capital, accountability, regulatory navigation) remains stubbornly human. Everything in between is contested territory, and the contestants are not other consultancies. They are agentic platforms.

Consulting firms are shifting from advisory-led models toward AI-enabled execution platforms, with leading firms embedding agentic AI, governed workflows, and repeatable digital assets directly into delivery environments rather than positioning AI as a standalone consulting service. Read that again slowly. The product of a consulting engagement in 2026 should increasingly be a living solution, a trained model or redesigned process, rather than a document about the solution. The slide deck, that imperial artefact of a hundred thousand airport hotels, is being quietly decommissioned.

What McKinsey's Arithmetic Means for Everyone Else

McKinsey has deployed thousands of AI agents through its proprietary Lilli platform, and its own executives have acknowledged that up to one-third of consultancy hours could be automated, potentially requiring staff retraining, changes in job scope, or workforce reductions. Let me translate that: a firm that built its identity on intellectual labour is telling its workforce that a third of what they do can be done cheaper, faster, and without the pension obligations.

If McKinsey, with its unrivalled brand moat, its century of institutional relationships, and its ability to recruit the top two percent of every MBA class, is running this calculation openly, what precisely is the value proposition of a firm incorporated last Tuesday with three directors and a Notion workspace?

Zero. Which is the number Companies House is recording.

Meanwhile, the Big Four and the execution-centric firms are growing at roughly double the rate of the pure strategy houses. Deloitte, EY, and Accenture offer both advisory and large-scale tech implementation. That combination, advice plus code plus accountability, is what enterprise clients are actually purchasing. Demand is strongest for consulting work tied directly to measurable operational and financial outcomes. Clients are pricing discovery workshops accordingly.

The Apprenticeship Gap Nobody Has Solved

There is a second-order problem that the AI evangelists at the large firms have not yet answered publicly, though they are certainly aware of it privately. The old consulting career path was learn by doing: juniors learned the fundamentals by sweating the details, fact-checking reports at two in the morning, double-checking calculations. Rote as it was, that grunt work was training. Now that AI handles much of it, firms have not yet established a new model for developing talent, and many insiders worry that future managers and partners will lack the same foundation.

The firms are automating the bottom of the pyramid at precisely the moment the pyramid depends on the bottom to train the top. This is not an HR problem. It is a long-cycle strategic risk that will manifest in partner quality roughly eight years from now, and nobody in a current partnership is especially incentivised to solve it.

The Proprietary Data Point That Changes the Framing

Return, then, to the Companies House number. Zero new SIC 70.22 incorporations in Q3 2026 is not merely a market signal. It is a forward indicator. Founders are rational actors. When the expected value of building a new management consultancy, factoring in competitive dynamics, AI-enabled client self-sufficiency, and the collapse of the billable-hour model, drops below the opportunity cost of doing something else, they stop incorporating. They are already doing something else.

This is the market delivering its verdict faster than any analyst report. Drucker observed that the purpose of business is to create a customer. Right now, the customer for traditional management consultancy is either a hyperscaler partnership, a vertical specialist, or increasingly, a subscription to something that does not require a registered company at all.

Prediction

By Q2 2027, at least two of the mid-market generalist consultancies with fewer than 500 staff will have repositioned as "AI implementation partners" or been quietly absorbed. The ones that survive will not be the ones who spent 2026 writing strategies about AI. They will be the ones who spent 2026 building them.

The register is not going to fill back up.