The idea has been around forever: software should be priced on the value it delivers, not on how many people log in. Easy to say, nearly impossible to implement. That's slowly changing.
Deloitte and Gartner both published reports this month pointing to the same shift: by 2030, more than 40% of enterprise SaaS spend will move to usage-, agent-, or outcome-based models. They're talking about large enterprise contracts for now. But waves don't stop at the enterprise shore.
To understand why seat-based pricing held for so long, you need to remember what it was solving.
People use software. They pay for licenses. Clean, predictable, easy to audit. The sales team knows what to quote. The finance team knows what to forecast. And the "value" of the product is tied to something tangible — the number of humans entering the system.
AI agents break this clean logic. An agent isn't a user. You can't assign it a seat. But it does the work.
Simple example: take a customer support platform. Five people, five licenses. Now an agent handles 200 tickets a day. The team is still there, but the work is different. Should the vendor still charge five seats? Or charge per ticket resolved?
Most vendors right now are quietly avoiding the question. That won't last.
Let me bring this to the domain I know best: clinical SaaS.
You're selling practice management software to dental or general medicine clinics. Core value: appointment scheduling, patient records, billing. You charge per practitioner. Now you ship an AI module. It improves appointment fill rates. It monitors patient treatment adherence. It sends automated follow-ups. A clinic using it increased their monthly revenue by 15%.
Fifteen percent more revenue. You're still charging the same monthly fee.
From the sales side, this is a great position to be in: your product is clearly more valuable, you have a strong argument for a price increase. From the customer side: "I'm paying extra for software that's monetizing my own patient data to improve their margins." That's a harder conversation.
The real appeal of outcome-based pricing is the sense of fairness. You pay when you get value. You don't pay when you don't. In theory, perfect alignment.
In practice, three stubborn problems.
First: measuring outcomes is hard. Isolating what the software actually caused is close to impossible. If a clinic's revenue went up 15%, how much of that was the AI module — and how much was a new associate joining, a marketing push, seasonal demand? Who measures, and how?
Second: planning and sales get harder. Customers want to know their costs in advance. "Pay as you grow" introduces revenue uncertainty on both sides. For small clinics, managing a variable software bill is a real operational headache.
Third: the trust problem. You say "we charge based on outcomes." The customer says "who measures the outcomes?" The answer is: you. That conflict of interest doesn't go away with good intentions.
So what's actually happening? Hybrid models.
A fixed base (platform fee or minimal seat charge) plus a variable component tied to measurable AI outputs. This keeps the sales motion simple, gives customers the "aligned incentives" message, and keeps the measurement dispute manageable — because the base is always there.
For clinical SaaS, this feels more realistic than full outcome-based pricing. A flat monthly platform fee regardless of usage, plus small per-action fees tied to specific AI outputs. Something like: 0.50€ per automatically filled appointment slot. A small charge per batch of automated patient reminders sent.
Small numbers individually. But they change the relationship.
There's something else going on in this conversation that I find more interesting than the pricing mechanics themselves.
When you shift — even partially — toward outcome-based pricing, you force a different internal conversation at the product company. You can no longer ship a feature and call it done. You have to ask: is this actually producing the outcome we're charging for? That question is uncomfortable. But it's the right question.
Most SaaS companies are better at shipping features than measuring their real effect on the customer's business. Outcome-based pricing, even as a light overlay on a base fee, forces that discipline.
And maybe that's the real reason it's hard to implement. Not the mechanics. The accountability.