Outcome pricing

Outcome pricing went mainstream in 2026. What buyers should take from it.

HubSpot, Zendesk and the big outsourcers all moved toward charging for results this year. Outcome pricing is only as good as the definition of the outcome and who checks it. Six questions to ask.

In 2026, paying for results stopped being a pitch and became a price list. HubSpot, Zendesk and some of the largest outsourcers moved at least part of their business to charging per outcome. Buyers now have room to ask for it. Outcome pricing is only worth having, though, if the outcome is defined in writing and someone other than the seller can check it was delivered.

What changed this year

Gartner put a number on the pressure in July: $234 billion of enterprise application software spend at risk from agentic AI, about 20% of enterprise SaaS spending by 2030. Its advice to software vendors was to "move from interface-based value to outcome-based value." In other words, stop charging for the screens people log into and start charging for the work that gets done.

Vendors moved in different ways:

  • HubSpot switched two Breeze agents to outcome pricing from April 14. The Customer Agent costs $0.50 per resolved conversation, down from $1.00, and the Prospecting Agent costs $1 per lead that is qualified and handed to the sales team.
  • Zendesk said at its Relate conference in May that customers would pay only for resolutions that are verifiably delivered. The agent confirms it has resolved the interaction, and a separate evaluation model then checks that claim.
  • Genpact said in August that revenue not priced per head (fixed fee, consumption and outcome contracts) had passed half its total for the first time, and that it expected more than $1 billion of agentic contract value in 2026.

Outsourcing buyers expect more of the same. In an HFS Research survey of more than 500 executives, outcome-based contracts were expected to nearly double from 20% to 39% of the mix within three years, while contracts priced per full-time employee fall from 42% to 28%.

Not everyone has moved. Microsoft still sells Copilot per user, and plenty of agent products are priced on credits or consumption. That makes it easier to compare: you can now put an outcome price next to a seat price and ask which one you would rather defend to your board.

The catch: who decides an outcome happened

An outcome price moves risk to the seller, which is the point. It also gives the seller a reason to count generously. A "resolved" conversation might be one where the customer stopped replying. A "qualified" lead might be one a model scored above a threshold.

There are three ways to verify an outcome, and they are not equal:

Who verifies How it works What to watch
The seller, on its own The agent or vendor reports that it succeeded You are trusting the party that gets paid
The seller, with a second check A separate model or process tests the claim before billing Better, but you still cannot see the evidence
A record you can audit Each billed item links to its inputs, steps and approvals You can check any line on the invoice

Zendesk's two-step check is a real improvement on self-reporting. The standard to aim for is the third: every invoice line tied to a record you can open.

Six questions to ask before you sign

  1. What exactly counts? Get a written definition for each outcome, including the evidence required and how long it must hold. An invoice posted and reversed a week later is not an outcome.
  2. Who verifies it, and can we see the evidence? Ask for a sample record for one billed item.
  3. What do we pay for failures? Agent time, retries, exceptions and items that fail their definition should cost nothing.
  4. What happens when volume spikes? Outcome spend rises with volume. Agree a cap.
  5. What does the floor fee cover? A fixed fee for running the service is reasonable. Make sure you know what is in it.
  6. If it is gain-share, what is the baseline? It should come from your own systems, agreed before go-live. See how to set a baseline for gain-share pricing.

How we price

We built Agentic MSP on the third model. Every agreement has a floor fee for running the service. On top of that, you pay per verified transaction or a capped share of verified gains. Each invoice line links to its record in the Evidence File. If the evidence is not there, the item is not billed. How our outcome pricing works.

Common questions.

What is outcome-based pricing for AI agents?

You pay for a defined result, such as a resolved customer conversation, a qualified lead or a posted invoice, rather than for user seats, licenses or the compute the agent used. The price is per outcome, or a share of a measured gain.

Which vendors moved to outcome pricing in 2026?

HubSpot moved its Customer Agent to $0.50 per resolved conversation and its Prospecting Agent to $1 per qualified lead from April 14, 2026. Zendesk said in May 2026 that customers would pay only for resolutions that are verifiably delivered. In outsourcing, Genpact said non-FTE revenue passed half of its total for the first time.

Is outcome pricing always cheaper?

No. It moves risk to the provider and ties cost to results, but spend rises with volume. Ask for a cap, and make sure failed items, retries and exceptions are not billed.

Sources

Want this run for you.

Agentic MSP runs back-office work with governed AI agents and bills only for verified outcomes.