Outcome-based pricing
A pricing model where the vendor charges per successful outcome — closed ticket, qualified lead, resolved bug — rather than per seat, per task, or per token. The signature 2026 agent pricing pattern.
Outcome-based pricing aligns vendor and buyer: the vendor only gets paid when the agent actually delivers the result. Sierra charges per resolved support conversation. Intercom Fin charges per resolution. 11x charges per qualified meeting. Decagon charges per deflected ticket.
It wins commercially because it short-circuits the procurement objection "we don't know if the agent will work." If the agent doesn't work, the buyer doesn't pay. Vendors that can stomach the variance get higher conversion and stickier contracts.
The complication is defining "outcome" precisely enough that both sides agree, including edge cases (partial resolution, escalated cases, false positives). Most 2026 outcome-priced contracts spend more legal cycles on the definition than the dollar amount.
Frequently asked
Outcome-based vs. per-task pricing — what is the difference?+
Per-task: you pay when the agent does work. Outcome-based: you pay only when the work succeeds. Outcome-based shifts execution risk from buyer to vendor; per-task does not.
When does outcome-based pricing fail?+
When "outcome" can't be measured cleanly (e.g., "good answer"). It works best where outcomes are binary and verifiable: ticket closed, lead qualified, PR merged.
