The disruption of professional services isn't coming. It's already happened in enough verticals to see the shape of it. The firms building the new model are operating, serving clients, and in some cases worth more than the entire industry imagined possible three years ago.
Law
Harvey AI was founded in 2022. By 2025, it was generating $100 million in annual recurring revenue and serving 42% of the Am Law 100. Its valuation: $8 billion.
NormAI, backed by Blackstone, operates with 35 lawyers and serves clients managing $30 trillion in assets. It is not augmenting a traditional firm. It is a different thing entirely.
Lawhive, backed by Google, acquired a traditional law firm — believed to be the first time an AI company has bought a conventional practice.
Among traditional firms that have already adopted AI widely, 20% report challenges meeting billable targets. The efficiency is real. The revenue model has not caught up.
Accounting
The Big Four collectively invested more than $10 billion in AI since 2023. PwC became OpenAI's largest enterprise customer. EY deployed 150 AI agents serving 80,000 tax professionals. KPMG launched its Workbench platform targeting $12 billion in added revenue.
The move is deliberate: AI lets the Big Four serve mid-market clients that previously wouldn't have justified their fee structures. Firms in the 20–200 employee range are their new target market.
At the same time, PwC cut 5,600 employees globally in the first half of 2025 and reduced U.S. graduate hiring by a third. The entry-level capacity is being replaced, not augmented.
The AICPA data shows what's on the other side of this transition: accounting firms with significant advisory revenue earn 30%+ higher monthly recurring revenue than firms focused on compliance work. The model exists. The question is which firms get there first.
Marketing agencies
S4 Capital — parent of Monks, one of the most prominent AI-forward agencies — reported H1 2025 net revenue down 12.7%, with headcount cut 8.9%. Its Technology Services segment fell 36.9%. S4's CEO: "The reality is AI is eating the agency business."
The Omnicom-IPG merger eliminated 4,000 positions with $750 million in expected cost savings — a defensive consolidation framed explicitly around AI.
83% of marketing leaders say they would reduce agency spending if they could fully automate content creation. 11% would stop using agencies entirely.
The agencies winning aren't fighting the automation. They're repositioning around judgment, strategy, and client relationships — and charging for those instead of deliverable volume.
Staffing
Bloomberg reported in February 2026 that AI threatens the $600+ billion global staffing industry as companies bring recruitment in-house. AI can now handle up to 80% of initial candidate sourcing and screening.
Unilever reviewed 250,000 candidates annually using AI without increasing recruiter headcount.
The firms adapting are moving from transaction-based placement fees toward workforce intelligence — continuous pipeline management, predictive hiring, skills gap analysis. The placement fee model is under pressure. The advisory model around talent is not.
Consulting
A 10-week project that can be executed in six is a 30–40% cost reduction. Clients know this. Pricing pressure is following.
McKinsey's internal Lilli AI assistant is used by 72% of its 45,000 employees, generating 500,000+ prompts per month — redirecting roughly $12 million per month in consultant capacity from research to client-facing work. That is the scale of reallocation happening inside one firm.
Ascentra Labs, founded by an ex-McKinsey team, is already adopted by three of the world's top five consulting firms. Early users report 2–4x faster execution on due diligence workflows.
The firms capturing value are the ones codifying their institutional knowledge into repeatable platforms — "service-as-software," in CB Insights' framing — rather than delivering bespoke projects.
What this adds up to
Every vertical has the same structure: AI-native or tech-enabled firms operating at a fraction of the headcount, backed by capital, taking share. Incumbent firms adapting their model. Firms doing neither watching margins compress.
The mid-market squeeze is real. Without the scale of global giants or the specialization of boutiques, firms in the 20–200 employee range are squeezed hardest. The firms that figure out what to sell differently — not just how to work faster — are the ones that end up on the right side of the split.
The data above is not a forecast. It is a current account.