Intercom's Fin AI agent charges $0.99 per resolution, billed only when a customer confirms the answer solved their problem or doesn't ask for more help. No resolution, no charge. Fin grew at an annualized 393% in Q1 2025 and passed about $100 million ARR. The company renamed itself Fin, and Salesforce agreed to buy it for about $3.6 billion in 2026, MarTech reported.
Key takeaways
- If your product does the work, charge for the work done, not the number of people using it.
- Outcome pricing is the easiest AI price to sell: you only pay when it works.
- It aligns the whole company. Every gain in resolution rate is more revenue.
- Define the outcome precisely and publish it. The definition is where disputes happen.
- Selling on top of rival tools means customers don't have to switch to buy.
Intercom had a problem.
It was a mature, respected help desk company. But growth had slowed to about 4% a year at its low point, by one pricing podcast's analysis. For a software company, that's the business equivalent of treading water.
Founder Eoghan McCabe came back as CEO. And he made a bet that most founders would find terrifying.
He put the company's future on an AI agent called Fin, and he priced it in a way almost nobody else was brave enough to try.
How Does Fin's 99-Cent Pricing Work?
Most help desk software charges per agent seat. You have 20 support staff, you pay for 20 seats.
Fin doesn't. It charges $0.99 per resolution, billed only when:
- the customer confirms Fin's answer solved their problem, or
- the customer doesn't ask for more help after Fin's last answer.
If Fin can't help, there's no charge. If the customer asks for a human, there's no charge. If the case is reopened, it's refunded. And Fin charges at most once per conversation.
Intercom later stretched the idea from "resolution" to "outcome." Completed workflows (called procedures) and successful handoffs now count too. Its pricing page (checked October 2026) also lists $1.99 per voice outcome and $9.99 per qualified sales lead.
Why Charge Per Resolution Instead of Per Seat?
Because Fin is designed to reduce the number of seats.
Think about it. If Fin works, a customer needs fewer human agents. If Intercom charged per seat, its best product would shrink its own revenue. That's a business model eating itself.
Charging per resolution flips it. Intercom gets paid for the work done, not the headcount.
In The Subscription Playbook, I said your usage metric "must be a metric that grows consistently with your customers and helps you communicate the value of your service." A resolved support ticket is about as clear a value metric as you'll find. Every customer understands it. Every customer can compare it with what a human-handled ticket costs them.
The Numbers: Did the Bet Pay Off?
| Date | Milestone |
|---|---|
| Low point | Company growth about 4% a year (secondary analysis) |
| Q1 2025 | Fin grows at an annualized rate of 393%; McCabe calls it "the largest quarter by net new ARR in the history of the company" |
| Jun 2025 | McCabe says Fin will pass $100 million ARR within about 2.5 quarters |
| 2025 | McCabe writes: "We're now at $400M ARR overall, with Fin about to pass $100M" |
| ~May 2026 | Intercom adopts the Fin name |
| 15 Jun 2026 | Salesforce agrees to buy Fin for about $3.6 billion (closed 10 Sep 2026), as reported by MarTech |
When a company renames itself after its new product, you know the bet worked.
What Did Outcome Pricing Change Inside Intercom?
This is the part I find most interesting.
Under seat pricing, the engineering team's work on answer quality doesn't show up directly in revenue. Under outcome pricing, it does. Every extra point of resolution rate is more money. According to Intercom, its R&D team now "fights for basis points."
That's alignment you can't buy with a mission statement. The customer wants more problems solved. Intercom earns more when more problems are solved. Everyone pulls in the same direction.
A help desk lesson, twice
In the book I told the story of Groove, another help desk company. It cut churn by 71% by spotting customers who weren't getting value and helping them before they left. Fin's pricing bakes that same idea into the bill: if the customer isn't getting value, Intercom doesn't get paid. You can't ignore your churn signals when they're your revenue.
Should you charge per seat, per conversation or per outcome?
Run your own numbers through the AI Agent Pricing Calculator and see which model earns more.
Why Did Selling on Rival Help Desks Matter?
Fin also works on top of competitors' help desks. A company using another tool can switch Fin on without moving its whole support operation.
That's a clever answer to the switching problem. In the MOATS framework, Switching is the moat that protects incumbents. Intercom went around its rivals' moats instead of trying to break through them.
Standalone Fin comes with a minimum monthly commitment (for example, 50 outcomes). That's Model #8 from the book, minimum commitment, doing its job: it guarantees a floor of revenue while the customer gets usage-based pricing above it.
How Does Fin Compare With Other Per-Resolution Prices?
Outcome pricing has become the standard way to price AI support agents. Here's how the main players compared in 2026.
| Vendor | Price per resolution (2026) |
|---|---|
| Fin (formerly Intercom) | $0.99 per outcome |
| Salesforce Help Agent | $2 per resolution |
| Zendesk | $1.50 to $2.00 per automated resolution |
| HubSpot | $0.50 per resolved conversation |
All of them price against the cost of a human-handled ticket. That's the anchor that makes the sale easy. See more in our outcome-based pricing examples.
When Doesn't Outcome Pricing Work?
Honestly? In most places outside customer support.
Outcome pricing needs an outcome you can count, that both sides agree on, and that happens often enough to produce steady revenue. A support ticket ticks all three boxes. "Better marketing" or "a happier team" doesn't.
It's also risky for the seller. If your product is mediocre, you earn very little. And your revenue swings with customer volumes you can't control.
What You Can Steal From Intercom
You don't need AI to price per result. Here's how a normal business could use this.
- Price per result where you can measure it. Per lead qualified, per booking made, per claim lodged, per issue fixed, per room cleaned.
- Write down exactly what counts. Intercom publishes its definition. Do the same, including what doesn't count and when you refund.
- Anchor against the alternative. Fin's price is easy to compare with a human ticket. What does your customer pay today to get the same result?
- Add a minimum commitment. A floor protects your cash flow while customers enjoy pay-per-result above it.
- Sell on top of what customers already use. If buying from you means ripping out another system, fewer people will buy.
Want to test the numbers? The AI Agent Pricing Calculator compares per-seat, per-conversation and per-outcome pricing for your own volumes. For the full method, read our guide to outcome-based pricing, then see how the new owner priced its own agents in Salesforce's Agentforce story.
Frequently asked questions
What counts as a resolution for Fin?
According to Intercom's pricing page (checked October 2026), an outcome counts when the customer confirms Fin resolved their issue, when they don't ask for more help after Fin's last answer, or when Fin completes a defined workflow (a procedure), including handoffs. Fin charges at most once per conversation, refunds reopened cases and charges nothing if the customer asks for a human.
Can you use Fin without Intercom's help desk?
Yes. Fin is sold standalone on top of other help desks, with a minimum monthly commitment (for example, 50 outcomes). That let Intercom sell to companies already using a competitor's software. It's a smart way to grow, because the customer gets the benefit without the pain of migrating their whole support operation.
Who owns Intercom now?
Intercom adopted the Fin name around May 2026. On 15 June 2026 Salesforce agreed to acquire Fin for about $3.6 billion, MarTech reported, and the deal closed on 10 September 2026. Fin's outcome-based pricing now sits alongside Salesforce's own Agentforce meters, which include a $2-per-resolution Help Agent.
Is $0.99 per resolution cheap?
It depends on what a human-handled ticket costs you. The price was designed to be easy to compare with that cost, which makes the return on investment obvious. Competitors price in the same range: Zendesk at $1.50 to $2.00 per automated resolution and HubSpot at $0.50 per resolved conversation, as of 2026.
What are the risks of outcome-based pricing for the vendor?
You only get paid when the product works, so a weak product earns very little. Revenue also depends on customer volumes you don't control. And you need airtight measurement, because customers will question what counted as an outcome. That's why Fin pairs outcomes with a minimum commitment and still sells seats for its help desk.
Get both books free
The Subscription Playbook (ebook + audiobook) and Feed a Starving Crowd, instantly. In Australia? Claim the free hardcover.
Robert Coorey, MBA
Robert is the author of The Subscription Playbook and Feed a Starving Crowd and a Director at BPAG Accountants in Sydney. More about Robert.
Sources
- Intercom, pricing page (checked 27 Oct 2026)
- Intercom Help, Fin AI Agent outcomes (checked Oct 2026)
- Stripe customer story, Intercom innovates outcome-based pricing for its Fin AI agent
- Eoghan McCabe on X (Jun 2025)
- Eoghan McCabe, There is exactly one way that SaaS can be saved (Fin Ideas, 2025)
- Kyle Poyar, Growth Unhinged, Intercom's decisive bet on AI
- Willingness to Pay podcast, Pricing Page unPacked: Intercom (secondary analysis of growth figures)
- MarTech, Salesforce acquires Fin, formerly known as Intercom (15 Jun 2026)
- Pricing Innovation, Agentforce's pricing journey (6 Oct 2026)
- EngagePulse, HubSpot Breeze pricing and credits explained (17 Jul 2026)
Facts and prices checked 27 October 2026. Spotted something out of date? Tell us.