AI Agent Pricing: Only 9% of Buyers Say Per-Seat Is Fair
EntrepreneurshipOnly 9% of B2B software buyers say per-seat pricing is the fairest way to charge for AI agents, the lowest of every option G2 tested, and many SaaS companies added agents to exactly that model last year. SaaStr’s breakdown of G2’s 2026 AI Agents Insight Report argues that founders who bundled an agent into an existing seat plan are heading into 2027 renewals on the model buyers rank last. For small SaaS teams, the data turns pricing from a positioning debate into a renewal risk.
The Number
G2 surveyed 1,019 B2B decision-makers in July 2026 for its 2026 AI Agents Insight Report. Preference for pay-as-you-go pricing, including consumption and per completed action or task, rose from 38% in 2025 to 57%. Pay-per-outcome rose from 11% to 19%. Fixed pricing fell from 35% to 21%. Per-seat came last, at 9%.
Adoption is broad: 86% of buyers say their company bought agent software or services in the past year, up from 69% in 2025, and 66% have agents in limited or scaled production.
How the Math Works
SaaStr’s explanation is simple. A seat counts the humans who log in, while an agent runs continuously and never logs in, so a per-seat price has little to do with the work an agent does.
Metering comes with its own problem. 46% of buyers report higher-than-expected operating costs, because agents use tokens, model calls and API calls on their own schedule and the bill arrives afterward. Buyers listed the controls they want before accepting variable spend, in order: real-time usage dashboards, monthly spend caps, cost-per-outcome reporting, routing to lower-cost models, and forecasting before deployment.
Oversight adds cost on the buyer side. 31% of organizations say employees spend six or more hours a week supervising agents, and 45% say oversight and rework consume more than a quarter of the productivity gains they expected.
What It Signals
Renewals are where this bites. 30% of buyers would consider switching to a vendor with stronger agent capabilities, and another 44% would pilot an alternative and reconsider at renewal, so SaaStr counts 74% as open to replacing a vendor. 72% call a strong prebuilt ecosystem of connectors, actions, skills or marketplace offerings very important or a must-have, and use of agent-builder platforms rose from 38% to 42% while ready-made agents fell from 30% to 27%. SaaStr’s reading: if your APIs, actions and MCP coverage are thin, buyers’ own agents can’t use your product.
Stack Archive laid out the alternatives in The 3 B2B Pricing Models Replacing Seat Licenses. G2’s data adds the buyer’s view of which ones they will accept, and on what conditions.
The So What
Founders charging for an agent inside a per-seat plan should split it out before their next renewal cycle: keep seats for human users, price agent actions or outcomes separately, and ship a usage dashboard and hard spend caps before any metered price goes live, since those are the first two controls buyers asked for. Six months before each large renewal, review measured agent results in that customer’s workflows with them, because three in four are prepared to run a competitive pilot.
Source: SaaStr
Content created with AI assistance and reviewed for accuracy.
Join the conversation
Stack Insiders is our free community for readers who want to go deeper — share resources, ask questions, and connect with others across every vertical we cover.
Join Stack Insiders →