SaaS pricing is a Ponzi scheme running out of new seats to sell

Per-seat pricing has quietly been the load-bearing wall of enterprise software for two decades. It's such a familiar mechanic that almost nobody questions the assumption baked into it: that the customer's headcount will keep growing, year over year, forever. That assumption just stopped being safe - and the entire revenue model built on top of it is more fragile than its valuation multiples suggest.

The math that worked until it didn't

Seat-based pricing is elegant for vendors because it auto-scales with the customer's growth without a single renegotiation. More employees, more logins, more revenue - no sales team required to capture the upside. It's also exactly why so much enterprise software optimized for adoption breadth (get everyone a login) instead of value depth (make the work itself better). Usage was a proxy for value because nobody had a cleaner way to measure it.

Agentic AI breaks that proxy in a way seat-based pricing was never designed to survive. When an agent absorbs work that used to require three logins, the vendor doesn't lose three logins' worth of value - the work still gets done, often better - but they do lose three logins' worth of revenue. The customer's headcount is shrinking or flattening precisely where vendors used to count on it growing, and no pricing committee built a model for that.

The renewal conversation nobody's ready for

Here's where it gets uncomfortable for incumbents: this isn't a future problem. It's showing up in next year's renewal conversations. A CFO who used to approve a seat-count increase without blinking is now asking why the seat count is flat while the agentic layer above the platform is doing more work than ever. Vendors have two options, and neither is comfortable - discount aggressively to defend the relationship, or hold the line and watch the renewal shrink anyway as the customer routes more work through agents instead of logins.

The vendors who say "we're adding AI features too" are answering the wrong question. Bolting an assistant onto a seat-priced product doesn't fix the unit economics. It just delays the renegotiation by a cycle or two.

What actually survives this

The replacement isn't complicated, just unfamiliar to legacy vendors: pricing tied to outcomes - tickets resolved, workflows completed, incidents closed - rather than logins provisioned. It's a model agent-native platforms can adopt naturally, because the unit of value they produce is a completed piece of work, not a seat filled. Outcome-based pricing is already showing up in agentic platforms like Airrived, where customers are billed for resolved work rather than headcount provisioned - a structural difference, not a marketing one, because it means the vendor's revenue grows when the customer's efficiency grows, instead of the two moving in opposite directions.

The next eighteen months of enterprise software renewals are going to be a referendum on which vendors saw this coming. The math was always going to catch up to the model. It just took an agent doing the work of three seats to make the gap impossible to ignore.

Anurag Gurtu is Co-Founder and CEO of Airrived, the Agentic AI platform recognized by Gartner for enabling enterprises to design, deploy, and govern autonomous agents at scale.