Liz’s view
The pre-AI economy ran on inputs. Lawyers billed by the hour, software companies charged by the seat. Effort stood in for value because value was too hard to measure. That’s about to change.
AI companies will want a direct piece of new business lines they unlock and the margins they fatten. These new pricing models will be a proxy for how the spoils of an AI-powered economy get divided.
“The arc of technology history will bend towards outcomes-based pricing,” Jake Saper, a partner at venture firm Emergence Capital, told me this week. Emergence was an early backer of Salesforce in 2003, just as software was shifting from upfront licensing agreements to cloud-based subscriptions. That move created a new billing unit — the per-seat credential replaced the fees to install software on your servers — and unleashed software’s golden age. A capital expense — a big one-time bill — turned into an operating expense with smaller ongoing bills. Now, it’s looking like it may become a value-share (no money down!), where AI companies get a portion of the money they make, or save, customers.
One of Emergence Capital’s bets is ProsperAI, a healthcare startup that verifies patient insurance before procedures. Most hospitals currently pay by the phone minute for humans to do this. Prosper does it with AI agents and is moving toward charging per successful authorization — outcomes, not inputs. “It’s teaching buyers how to buy,” Saper said.
Investment-banking startup Rogo charges the likes of Goldman Sachs and JPMorgan per user today but says it’s exploring ways to take a cut of the M&A and IPO revenues it helps generate. As AI makes it profitable to chase smaller deals that were previously uneconomic, “it’s upside for them and it’s upside for us,” Rogo president Rahul Rekhi said.
AI is even coming for the last great holdout of input-based pricing: Big Law’s billable hour. “Clients want value, they are willing to pay for value, and the way that value is measured is ripe for evolution,” said Rachel Proffitt, CEO of Cooley, the global law firm headquartered in Silicon Valley.
The legal industry’s future may look a lot like its oldest-school heavyweight, Wachtell Lipton, which famously sends clients bills that simply say “For Services Rendered” next to a large number. You hire Wachtell for an outcome — a gold-plated M&A deal — and pay accordingly.
Economic units matter because they set incentives. When Rolls-Royce’s “power by the hour” model, introduced in the 1960s, began charging airlines for engine operation time rather than parts upfront, it forced Rolls-Royce to internalize maintenance costs, and planes got more reliable. The media industry’s shift from charging for eyeballs to driving subscriptions changed the incentives from clickbait to quality journalism.
AI companies sharing in the new revenue they create may siphon money away from powerful law firms and financial shops, or at least allow the giants of those fields to put smaller rivals out of business. It will almost certainly cost jobs. But aligning service providers’ incentives with their customers is the right move.
Notable
- AI still has a long way to go before it starts thinking like a lawyer — and it won’t replace them anytime soon, AI reporter Michelle Kim argues in MIT Technology Review.