The Cathedrals Are Hollow

The Big Four—Deloitte, PwC, EY, KPMG—plus McKinsey, BCG, and Bain have spent $10 billion on AI since 2023. The architecture is magnificent. The interiors are empty. A consultation that cost $800k in 2023 now takes one person and a Claude prompt. Clients still walk away with a slide deck and a "roadmap." The engine room hasn't moved.

I was one of fifteen Innovation Scouts in a 55,000-person insurance organization. Hartford Insurance, Munich Re. We identified inefficiencies that cost the organization tens of millions annually. Almost nothing changed. The manuals were in place. The procedure was written. The institutional mass resisted the small currents of change with the force of geology. I watched this inside one firm. The Big Four are doing it at planetary scale. Spending like change is coming, acting like it isn't.

That is the hollow cathedral: imposing from the outside, structurally vacant inside.

The Numbers That Hurt

Ninety-five percent of PE funds report AI initiatives meeting or exceeding their original business case criteria. But only 17 percent report *significantly* exceeding them. The gap between "meets expectations" and "transforms the business" is a chasm. It reveals the real story: AI is compressing engagements, not reinventing them.

BCG reported that 25 percent of its $14.4 billion in revenue. $3.6 billion. Came directly from AI consulting work in 2025. That is real growth. But it is growth in a broken model. The revenue numbers look right. The delivery method is unchanged. Outcomes-based pricing now accounts for roughly 25 percent of McKinsey's fees. That same McKinsey still bills 75 percent of its work the old way.

The US AI consulting market will exceed $15 billion in 2026. Accenture alone owns 77,000 AI professionals. The cathedrals are still being built. The congregation is leaving.

Why the Pyramid Model Persists

A consulting engagement works like this: senior partner sells the engagement, manager oversees the work, senior consultant designs the solution, consultant executes it, analyst gathers the data. Each layer makes margin on the layers below. Each layer justifies its existence by billable hours.

AI collapses the analyst layer first. The routine work. Data gathering, report formatting, initial analysis. Vanishes into a large language model. Firms respond by hiring fewer analysts. Deloitte cut its UK graduate intake by 18 percent. EY cut by 11 percent. PwC cut by 6 percent.

But the senior partner still needs to sell. The manager still needs to oversee. The senior consultant still needs to show up. AI didn't destroy the pyramid. It just hollowed out the base. The structure stands. The bottleneck moved up.

Competence beats credentials. An owner-operator consultant who can run end-to-end on a problem, who owns the outcome, who doesn't need three layers of reporting. That person will outrun the pyramid every time. But the pyramid rewards the layers, not the outcome.

What Clients Are Saying (But Not Out Loud)

Eighty-six percent of consulting buyers actively seek AI-enabled services. Sixty-six percent said they would stop working with firms that fail to incorporate AI. Clients are speaking clearly: we want speed and smarter work, not more junior heads.

Yet 70 to 85 percent of AI initiatives fail to meet expected outcomes. Forty-two percent of companies abandoned most AI initiatives in 2025. The gap between what clients demand and what they receive is the architecture of the hollow cathedral. The firms have the tools. They do not have the system. A Harvard Business School study of 758 BCG consultants found AI users completed 12.2 percent more tasks, 25.1 percent faster, with over 40 percent higher quality output. These are individual consultants with better tools, not transformed delivery models.

When hybrid teams combine human consultants with AI systems intentionally. Not as bolt-on acceleration, but as redesigned engagement. They deliver projects 35 percent faster while maintaining quality. That is the exit. That is the acquirable model. The consulting firms know it exists. They have not built it at scale.

The Shift That Hasn't Happened Yet

Outcomes-based pricing should be the natural correction. If you deliver faster, you charge less per day but more per outcome. The margin stays on the solution, not on the clock. Sixty-seven percent of consulting buyers now prefer fixed-fee arrangements over time-and-materials contracts. That is only three years of change. The demand is real.

But the supply side has not followed. A project that once required a six-person team for three weeks now requires one person and a few prompts. The firm still wants six people. The client still wants to pay for three weeks. Both are wrong, and both know it. No one is willing to take the first loss. When one firm renegotiates outcome-based contracts at compressed margins, it sets precedent that compresses margins industry-wide. Each firm would rather hold price and lose the deal than renegotiate downward and poison the market.

This is how competence stays beaten by credentials. The system protects the credential. The system does not protect the outcome.

The Exit Is Owned, Not Sold

An owner-operator in consulting would be unrecognizable to the pyramid. She would own a vertical or an industry. She would take accountability for the outcome, not for billable hours. She would hire people who could execute end-to-end, not people who fit a staffing model. She would price based on the result, not on the input. She would use AI to amplify her own competence, not to justify layers of management.

That person would run circles around the Big Four. She would be acquirable because she would have a system, not a hierarchy. She would be sellable because she would have compounding revenue, not rotating engagements.

The $10 billion spend on AI was an investment in the cathedrals, not in competence. The firms bought tools. They did not rebuild the foundation. The manual is still the pyramid. The procedure is still billable hours. The sovereign model. The one that survives. Belongs to the owner-operator.

Doctrine says it first: competence beats credentials. The Big Four have credentials. They do not yet have competence in the new system. When they do, they will no longer be Big Four. They will be smaller and faster, and they will still lose to the people who started believing it in 2026.

FAQ

Q: Are consulting firms actually abandoning the pyramid model?

A: Not yet. Hiring is mixed. McKinsey plans to hire 12 percent more graduates in 2026 than this year, signaling confidence in traditional pyramid scaling. But EY, Deloitte, and PwC are cutting entry-level hiring, suggesting they're preparing for a hybrid model. Most firms are in transition, not transformation. They're cutting the base while keeping the structure intact.

Q: If 95% of PE funds say AI meets expectations, isn't that success?

A: It's a ceiling disguised as a floor. Meets expectations means "we got what we paid for." Only 17% report significantly exceeding expectations. The AI initiatives are working; the business model isn't changing. Clients are satisfied with incremental gains when they expected systemic change. That gap is where the hollow cathedral lives.

Q: Why haven't outcome-based pricing models taken over yet?

A: Margin compression. A six-week engagement billed at $400k per week generates $2.4 million. AI-compressed to two weeks still needs to justify the same partner economics. Someone takes a loss on the new model. The market is waiting for the first firm to break. When they do, everyone else gets margin-pressured. Until then, the old pricing holds.

Q: What would an owner-operator consulting firm actually look like?

A: A vertical-focused practice led by someone who owns the P&L end-to-end. She hires consultants who can do senior work, not layers. She prices outcomes, not hours. She uses AI to amplify her capability, not to justify staffing. She exits to a strategic buyer at a 10x multiple because she has a system, not a rolodex. She doesn't exist at McKinsey. She exists in the gaps the Big Four left behind.

Q: Is the consulting industry broken or just mid-transformation?

A: It's mid-transformation that looks like stability. The cathedrals still stand. The congregation still attends. But the ritual doesn't match the structure anymore. In three years, the firms that adapted. That made the pyramid flatter, the pricing outcome-based, the outcomes measurable. Will have compounded growth. The ones that didn't will have higher margins today and smaller pipelines tomorrow.

*Jeff Barnes is the founder of Digital Evolution Marketing Group (DEMG). demg.ai has no commercial relationship with any vendor, platform, or tool mentioned in this article. This content is for educational purposes only and does not constitute business, legal, or financial advice. Results described are illustrative and may not reflect your specific situation.*