AI's Double-Whammy: Who's Actually Making Money?

Every enterprise conversation about AI right now is about adoption and ROI, spanning pilots to production, models and tokens, value realization. It's a demand-and-growth story, and it sounds promising.

Look at the other side of the table, inside the providers building and delivering AI, and a nervous story is playing out. The supply side is in upheaval. And pricing sits right at the center of it. Here's why: for the last two years, frontier AI has been priced artificially low, heavily subsidized by cheap capital chasing adoption and market share. That's ending, because it was never sustainable. Providers (services firms, SIs, consultants, AI-native products, AI-augmented platforms) are now facing the real economics of what they're selling, often for the first time. And almost everyone we've talked to is improvising.

Value realization is already the hard problem as enterprises struggle to prove AI's ROI, and providers get scrutinized hard for it. Now layer in the cost problem: the changing and increasing nature of cost. That is the double whammy, a highly challenging problem for the providers to solve.

We've heard the same handful of questions, in different words, from very different companies:

• If model costs keep falling, should we cut our prices or hold them and pocket the margin? Most remain undecided.
  
• When AI frees up an hour of a delivery team's time, where does that hour go? Very few are tracking it well enough to say.
• Is per-seat pricing going to survive, or is usage-based or outcome-based pricing inevitable, and if so, on what terms?
• How much of the productivity everyone's claiming is turning into ROI and contributing to margin?
• Who inside the org even owns AI economics (cost, pricing, value, margin together)?

It's early, and it's understandable that the playbook doesn't exist yet. But that means most providers are running blind right now and therefore, setting prices, structuring deals, and making margin bets on instinct, competitor-watching, and hope.

We think that's a problem worth solving with data, not opinion or guesswork.

So we're running a study to answer it. From the inside, sourced directly from the people who own these decisions across the market: services and BPO leaders, SI and consulting partners, AI-native product teams, and AI-augmented platform teams. Four vantage points, one common (also customized) set of questions. So the patterns that cut across the market become visible for the first time.

We're not building this to confirm or validate what everyone already assumes. We're also building it to find the aha’s, so strategy can be built on real, collective intel from the ground, not guesswork.

If any of this sounds familiar: if you've argued about whether to hold your price or cut it, if a buyer showed up with a benchmark you couldn't counter, if you've claimed productivity gains you couldn't actually trace to the bottom line, we built this for you. This research exists because of conversations exactly like that.

Give us a shout if this is the conversation you're having internally. More on this to follow soon.