SaaS vendors are rethinking how they charge for software as AI features move from add-on to core product, a shift Fast Company examines in its report "How AI is rewriting the rules of SaaS pricing."

The core tension is straightforward. AI workloads cost money every time a customer uses them — each query can trigger inference costs that never existed in the per-seat pricing model most SaaS companies were built on. A traditional seat license scales with headcount; AI costs scale with usage. That mismatch is what's forcing the pricing conversation.

Why does per-seat pricing break under AI?

For two decades, the standard SaaS contract charged per user per month. That worked because one more login cost the vendor essentially nothing. AI changes the math. When a feature consumes compute on every call, a customer with 10 power users can cost more to serve than one with 100 casual users — even though the second contract pays ten times the revenue.

Vendors responding to this are moving toward models that track what the software actually consumes:

• Usage-based pricing, billed on API calls, tokens, or transactions
• Hybrid structures: a base platform fee plus usage overages
• Outcome- or value-tied pricing, where charges track results the AI delivers

Each carries tradeoffs. Usage pricing aligns cost with value but makes revenue harder to forecast for both sides. Hybrid models protect a revenue floor but can confuse buyers. Outcome pricing is the most defensible in theory and the hardest to measure in practice.

What should product managers watch?

Pricing is a product decision, not just a finance one. Teams shipping AI features need instrumentation from day one: what a unit of usage costs to serve, how customers actually consume it, and where the margin cliff sits. Without that telemetry, any pricing model is a guess.

The failure modes are real. Mispriced AI features can quietly erode gross margin as adoption grows — the better the product, the worse the economics. On the buyer side, unpredictable usage bills push procurement teams back toward flat contracts, stalling expansion.

The direction of travel is clear: as AI becomes table stakes in SaaS, pricing will keep shifting from seats to consumption, and the companies that instrument usage early will set the terms for everyone else.

Note: The source item provided only the headline; specifics above reflect the well-established parameters of the SaaS pricing debate it points to.