Enterprise Software Pricing Strains Under the Weight of AI
Per-seat models built for human users fit poorly when software does the work, and vendors are experimenting their way toward outcome-based alternatives.
.jpg?width=800)
clender via Wikimedia Commons · CC BY 2.0The per-seat license survived every previous platform shift. It is not surviving this one intact. Procurement teams renegotiating enterprise agreements this year describe pricing conversations that open with a question no sales playbook anticipated: how many of the seats on this contract are still occupied by people?
The candidates are proliferating: consumption pricing tied to tasks completed, hybrid models that meter AI usage atop a platform fee, and outcome-based contracts that charge per resolved ticket or processed claim. Each shifts risk between vendor and customer differently, and procurement teams report that no two proposals look alike.
On the buy side, the chief complaint is forecastability. Consumption pricing without caps turns annual software budgets into guesswork, and outcome definitions invite dispute at renewal. The contracts gaining traction pair usage-based economics with committed floors and negotiated ceilings, familiar territory imported from cloud agreements.
For vendors the stakes are existential arithmetic. If AI lets a customer accomplish the same work with fewer seats, revenue models tied to headcount shrink with their customers' efficiency. The companies moving earliest to decouple price from people are, not coincidentally, the ones most confident their software will replace some.