Here's the trap. Outcome pricing sounds like the fairest deal in software: you only pay when the thing works. So you skip the definition and start modelling volumes. The definition was the product.
Intercom Fin charges $0.99 per outcome, read off its own pricing page on 29 August 2026. One of the three things that counts as an outcome is the customer not asking for more help after Fin responds.
That is what a satisfied customer looks like. It is also what somebody who gave up and closed the tab looks like. From the billing system's point of view they are the same event.
So this page gives the verified rate, the definition in full, what the other vendors in the inbound qualification category publish, and what to ask before you model any of it.
What does Fin actually charge?
$0.99 per outcome. Worth noting that the word has changed: the category, and our own records until today, called this per-resolution pricing. Fin's live page says outcome. Same meter, softer noun.
Its published definition, in full:
Read that last sentence first, because it is the genuinely good part and it deserves credit. One charge per conversation, however many questions get answered, is the opposite of a per-message meter and it protects you from exactly the runaway most usage pricing produces.
Why is the second trigger the important one?
Because it bills silence. If a customer stops replying, that counts as an outcome, and a customer stops replying for two very different reasons. Either the answer worked, or they concluded they were not going to get one.
Nothing in the definition separates those. There is no requirement for a positive signal, no satisfaction check, no minimum engagement. The absence of a follow-up question is the billable event, and abandonment produces the absence of a follow-up question just as reliably as success does.
This is not an accusation. Fin publishes the rule plainly on a public page, which is more than most of this category does, and publishing it is the transparency. The point is that a buyer has to read it, because the metric you are paying against is not the metric you think you are buying.
What about the third trigger?
It reads: Fin completes a workflow, including handoffs. So the agent can take a conversation, run its procedure, decide it needs a human, pass it over, and charge 99 cents for doing so.
In fairness, routing correctly is genuine work and a good handoff beats a bad answer. But it is worth being clear about what you are buying: in that scenario you pay the software and your team still handles the conversation. If a large share of your traffic ends in handoffs, outcome pricing is closer to a per-conversation fee with extra steps.
What do the other vendors charge?
Mostly, they will not say. Sierra prices per resolution and publishes no rate. Decagon is custom. Crescendo prices per resolution and publishes no rate either. Across this category of eighteen tools, sixteen publish no price at all, which makes Fin the outlier for publishing anything.
The self-serve options that do publish are a different product class rather than a cheaper version of the same one. Chatbase at $40 and Landbot at $45 are both real prices, verified in August 2026, and neither is an autonomous agent. Our own notes are blunt: Chatbase answers questions but does not act on systems, and Landbot is a form builder with conversation on top.
So what should you ask before signing?
- Ask for the billing definition in writing, from every vendor. Fin's is public. For the fourteen or so that publish nothing, this is the single most consequential sentence in the contract and it will not be volunteered.
- Ask what share of conversations end in a handoff, and whether those are billed. At Fin they are. If your traffic is complex, that ratio determines whether outcome pricing is cheap or is a conversation fee wearing a nicer name.
- Ask whether abandonment is distinguishable from resolution in the reporting, even if it is not distinguishable in the billing. If the vendor can tell you which conversations ended in silence, you can at least audit what you paid for.
- Model at a rate rather than a volume. Nobody, including us, can tell you how many outcomes a month you will have. What you can do is take your current conversation volume, assume most of them bill, and see whether 99 cents each is a number you would accept as a flat per-conversation price. If not, the outcome framing is not going to save you.
- Check whether you can buy it at all this quarter. Two of the three enterprise options have no self-serve path, which puts them in the majority of tools that require a sales call before they will show you a number.
The answer to the question in the title is that at Fin an outcome is a confirmation, a silence, or a handoff, and two of those three are things that happen when the software did not solve the problem. That does not make it a bad deal at 99 cents. It makes outcome pricing a different thing from what the phrase implies, and the vendors who publish no definition at all are asking you to accept the phrase without it.