The Big Four and top strategy houses have poured over $10 billion into AI since 2023. Almost all of it went inward.
PwC became OpenAI's largest enterprise customer. KPMG signed a $2 billion Microsoft alliance. Firms equipped their own analysts with knowledge tools and compressed internal research timelines from weeks to days. The productivity gains were real. They just didn't reach the client.
A task that used to take a junior team three weeks now takes three days. The client still gets invoiced for the original timeline.
This isn't a conspiracy. It's a structural incentive problem. When your revenue model is built on hours and headcount, every efficiency gain is a margin improvement for the firm, not a cost reduction for the client. There's no mechanism to pass the value through.
And clients are starting to notice.
The $172 million signal
In late 2025, medical device manufacturer Zimmer Biomet sued Deloitte for $172 million over a failed software implementation. The case is still being litigated, but the framing is the point: a client paid a major firm to deliver a working system, didn't get one, and is seeking direct financial recovery.
That's not how consulting disputes typically end. Usually they resolve quietly — scope disagreements, change orders, extended timelines. A $172 million public lawsuit over delivery failure signals a different posture from the buyer side.
The Zimmer Biomet case isn't isolated. It's the most visible example of a pattern building across industries. Clients are asking harder questions: If AI can reduce your analysis timeline from weeks to days, why hasn't my fee changed? If your team is half the size it was, why is my bill the same?
AlphaSense's recent consulting industry analysis found that clients no longer want recommendations — they want partners who implement solutions and drive measurable results at speed. A Deltek survey found that 73% of clients now expect real-time visibility into project status. Transparency has shifted from differentiator to baseline.
Some enterprises are bypassing consulting firms altogether. Tech-savvy companies are building internal AI capabilities rather than paying for external teams. The logic is straightforward: if an off-the-shelf AI tool and two internal analysts can match what a consulting team did, faster and cheaper, why wouldn't you?
The boutique advantage — with an asterisk
This shift isn't automatically a win for smaller firms. Boutique consultancies that still bill by the hour face the same structural problem as the Big Four. They just have less overhead absorbing the mismatch.
The firms gaining traction are the ones that restructured around outcomes. Fixed scope, fixed fee, defined deliverable — where the contract specifies what gets built, what business result it ties to, and who carries the execution risk.
The most compelling AI businesses don't sell tools or time. They sell the work.
— Sequoia Capital, "Services are the new software"
The incentives flip completely. When your fee is tied to a defined result, every AI workflow you build and every automation that compresses a timeline improves your margin and delivers more value to the client simultaneously. Interests align instead of compete.
We run our own firm this way, and the operational difference is measurable. An engagement that would require a four-person team and six weeks under an hourly model often gets delivered by one operator and an AI-native stack — n8n, Airtable, Claude, and a few purpose-built agents — in under three. The client buys the result for the same price either way. The firm keeps the efficiency gain. Neither side subsidizes the other.
Three questions buyers should ask
If you're evaluating an AI engagement in 2026, three questions separate the models.
Is the fee tied to a defined deliverable and business outcome, or to time and team size?
If it's the latter, you're absorbing the provider's efficiency risk — and funding their AI investment with your budget.
Does the engagement end with a working system you can operate, or with a document that describes one?
A trained model, a redesigned process, an automated workflow — these are what delivery looks like now. Strategy decks that don't compile into operations are increasingly hard to defend.
Who carries the risk if the project doesn't deliver?
In an outcome-based model, the provider has a contractual stake in the result. In time-and-materials, the only guarantee is that the team showed up.
Where this is heading
The consulting industry isn't going away. The expertise is real, and the need for external perspective isn't shrinking. What's changing is how that expertise gets packaged, delivered, and priced.
The firms that will define the next decade are the ones restructuring around what clients actually need: defined outcomes, delivered efficiently, with commercial models that align provider incentives with client results. The firms still selling the old pyramid — expensive teams, hourly billing, strategy decks — will find themselves competing against internal teams, AI tools, and outcome-based providers simultaneously.
The Zimmer Biomet lawsuit isn't the end of consulting. It's the end of consulting that can't show what it delivered.
We sell the work, not the tool. That distinction used to be positioning. In 2026, it's becoming the only sustainable business model.
At Tellefsen, every engagement is structured around fixed outcomes and AI-native delivery. If you're exploring what an outcome-based AI engagement looks like, we'd welcome that conversation.
Book a conversation