Last week, two of the Big Four signed major firm-wide deals with Anthropic in the span of about seven days. PwC announced it’s training and certifying 30,000 professionals on Claude. KPMG announced its Digital Gateway — embedding Claude into its core delivery platform for all 276,000-plus global employees. Deloitte already has an existing Anthropic relationship. EY is the only one of the four that hasn’t made a comparable announcement yet.
I’ve been working with midsize firms on AI rollouts, and this week’s news crystallized a question I keep coming back to: at what point does everyone having the same tool stop being a competitive advantage and become just the price of admission?
We’ve Been Here Before
Technology commoditization is a recurring arc in professional services. Think about audit software in the late 1990s and early 2000s. Early adopters had a real edge — they were faster, more consistent, better documented. But by the mid-2000s, every firm had it. Nobody was winning business because they used audit software. It became the floor, not the ceiling.
Same thing with cloud and SaaS ERP. Being “on the cloud” was a differentiator in 2014. The same arc played out over the decade. Today nobody leads a client pitch with “our people use cloud software.”
AI is going through the same cycle — except the timeline is compressed. If you’d said two years ago that you had an AI-enabled delivery platform and could prove it, you’d be having a very different conversation. You’d be a differentiator. If you say that today, most people would not be surprised. Some would be more surprised if you said you weren’t using AI at all.
Two Things Are Happening Simultaneously
The first is that the supply side is moving in lockstep. When PwC, KPMG, and Deloitte are all building on the same underlying model, the technology layer becomes neutral. It’s a shared input — not a differentiator. Whatever advantage exists has to come from what you build on top of it: the specific workflows, the quality of judgment and data, the institutional knowledge you can capture, and the depth of the client relationships underneath it all.
The second is that the cost is real and significant. These deals start looking approachable — $20 flat access fees are how most people got anchored. But enterprise token-based pricing can run five times that or more for the same volume of use. A lot of leadership teams are writing large checks before clearly answering what differentiated value they’re actually buying. The honest internal answer I hear most often: “We’re not sure yet — but we can’t afford not to.”
Defensive Play vs. Offensive Play
Here’s the framing I think is most useful. Because of the nature of general-purpose AI tools and how broadly they’re being adopted, most AI investments right now are going to be defensive plays, not offensive plays. The problem is that a lot of firms have them framed as cheap offensive plays — when they might actually be very expensive defensive plays.
That distinction matters for how you budget, how you sell it internally, and how you set expectations with leadership. A defensive investment protects your position. An offensive one creates new ground. Both are legitimate — but they’re not the same thing, and pretending one is the other creates a credibility gap when the outcomes don’t match the pitch.
Two Frames for Evaluating AI Spend
If you’re a practitioner at a firm: Stop asking “should we adopt AI?” Start asking “what specifically will we do with it that our competitors can’t replicate in 90 days?” If the answer is roughly what KPMG is doing — document review, research, drafting — you’re buying parity. That’s still worth doing, but budget for it as an operating cost, not a transformation program. Don’t pitch it internally as a leap forward if it’s really a catch-up move.
If you’re advising clients: Help them pressure-test vendor AI claims with one question: is this genuinely changing the outcome I’m getting, or is everyone now offering the same thing with different branding? If a Big Four team and a regional firm are both running the same model on the same type of engagement, the fee differential needs some explaining.
Key Takeaways
- When every major firm uses the same AI, the technology stops being the differentiator. Advantage lives in the workflow, judgment, data, and client relationships built on top of it.
- AI adoption at scale is increasingly a parity cost — necessary to stay competitive, but not a transformation story. Firms that pitch it as a leap forward when it’s really catch-up are setting themselves up for a credibility problem.
- Most AI investments right now are defensive plays. Framing them as offensive plays creates misaligned expectations around cost, ROI, and outcomes.
- The 90-day replication test: if your top competitor can replicate what you’re doing with AI in 90 days, you’re buying parity. If they can’t — figure out why, and protect that use case.
Want the CPE credit? Take the full lesson on EverydayCPE and earn 0.2 CPE credits: [lesson link]
Leave a Reply