Only 9% of Buyers Call Seat Pricing Fair for AI Agents

Only 9% of B2B buyers call per-seat pricing the fairest model for AI agents, and 74% are ready to shop you at renewal. Both numbers come from G2’s 2026 AI Agents Insight Report, a July 2026 survey of 1,019 B2B decision-makers, as analyzed by Jason Lemkin on SaaStr.

The pricing preference is moving fast. Preference for pay-as-you-go rose from 38% to 57%, and pay-per-outcome rose from 11% to 19%. The renewal figure breaks into two parts: 30% would consider switching to a vendor with stronger agents, and another 44% would pilot an alternative and revisit at renewal. Buying is not the constraint. Some 86% of buyers purchased agent software in the past year, up from 69%, and 66% have agents in limited or scaled production.

Cost surprises explain much of the pressure. Forty-six percent report higher-than-expected operating costs, and buyers want usage dashboards and spend caps. The labor claim is also thinner than many pitches suggest: 31% say staff spend six or more hours a week supervising agents, and only 14% report a net headcount reduction. A seller promising pure replacement is walking into a buyer who has already measured the supervision time.

Lemkin’s advice is practical. Run competitive pilots in top-20 accounts six months before renewal. Price agent work separately from seats. Ship spend caps before metered pricing. Add cost-per-outcome reporting. Prioritize connectors and MCP support in 2027.

For a seller, the lesson is to treat renewal as a sales cycle that starts half a year out. If your product includes agents, prepare to show usage, cost and outcome data to the buyer who now expects it, and to defend your price against a pilot that a competitor is already proposing. The survey is a snapshot of buyer intent, not proof of what they will do, but the direction is consistent.