Overview
For two decades, SaaS meant a predictable formula: charge per user, per month, for access to a tool humans operate. AI strains every part of that sentence. This report explores what the model is becoming.
Why per-seat pricing weakens
Per-seat pricing assumes humans do the work and you sell each human a seat. But if AI agents do much of the work, a company needs fewer human seats to get the same output — directly shrinking seat-based revenue. The pricing model that built SaaS is misaligned with a world where software, not people, does the doing.
The rise of usage and outcome pricing
Two replacements are emerging. Usage-based pricing charges for what's consumed (API calls, generations, tasks), aligning revenue with both value delivered and AI cost incurred. Outcome-based pricing goes further — charging for results (tickets resolved, leads booked, code shipped) rather than access. Buyers increasingly want to pay for outcomes, not logins.
From tools to workers
The deeper shift: software is moving from a tool you operate to a worker that does the job. When you sell "the work done" instead of "a tool to do the work," you're competing with labor budgets, not software budgets — a much larger pool, but with different expectations around reliability and accountability.
SaaS mutates, not dies
This isn't the death of recurring software revenue — it's its evolution. The companies that thrive blend software and service: software that delivers outcomes, priced by usage or results, with agents doing more of the labor. "Software as a service" becomes, increasingly, "service, delivered by software."
What this means for you
If you build software: revisit pricing — can you align it to usage or outcomes? Can your product do the work, not just enable it? If you buy: expect more outcome-based options and value them. If you invest: watch which companies adapt pricing and which cling to per-seat in an agent world.
Honest limits
Per-seat SaaS won't vanish soon — it's simple to buy and budget, and many tools still augment humans rather than replace them. The shift is directional and uneven, strongest where agents can reliably own outcomes.
