Intercom Fin AI Agent Pricing: Fin AI Cost per Resolution and Outcome Rates

Fin costs $0.99 per outcome, and unlike most of its competitors Intercom publishes that number where anyone can read it. An outcome is counted when a customer confirms their issue is resolved, when they stop asking for help after Fin responds, or when Fin completes a Procedure, including a handoff. Resolutions, Procedure handoffs and disqualifications all bill at $0.99. Lead qualifications bill at $9.99.

Three rules decide what actually lands on the invoice, and all three are in Intercom's favor. You are charged once per conversation however many tool calls, lookups and replies Fin makes inside it. You are not charged at all when Fin passes a conversation to your team without reaching an outcome. And there are no setup, integration or platform fees, which is genuinely unusual in this category and worth checking for on every competing quote.

The only fixed cost is a minimum of 50 outcomes a month. That puts a floor of $49.50 a month under any Fin deployment, which is small enough that it will not decide anything, but it does mean Fin is not free in a month where nobody contacts you.

This is where most Fin forecasts go wrong. Support teams count tickets. Intercom bills resolutions. Multiply the wrong one by $0.99 and you will be out by whatever your resolution rate is, which is not a rounding error.

Do the conversion once and keep it. Cost per inbound conversation equals $0.99 multiplied by your resolution rate. Intercom publishes a fleet average of 76% across more than 12,000 customers, with some accounts above 85%, and says the average rate climbs about one point a month. Third party analyses of Intercom's own published case studies land lower, in the 42% to 50% range, which is the sensible number to plan a first year on. Both are on the table below, because the honest answer is that your rate depends on your help center and nobody can quote it for you.

Read the shape of that table before the numbers. A better agent costs you more. Going from a 42% resolution rate to 76% roughly doubles your Fin bill on identical ticket volume, and every one of those extra dollars bought a conversation a human did not have to touch. That is outcome pricing working exactly as designed, and it is still the sentence most likely to surprise a CFO who approved the pilot on the low number.

On the same rate card, in the same list, a lead qualification costs $9.99 against $0.99 for a resolution. That is just over ten times the price, and it is the difference between a support deployment and a sales deployment.

The reason it matters is that Fin is sold for service, sales and ecommerce on every Intercom plan, so pointing it at your pricing page to qualify inbound leads is a normal thing to do and takes about an afternoon. Run 500 qualifications a month at $9.99 and you are at $4,995. Run 500 support resolutions at $0.99 and you are at $495. Same volume, same product, same invoice line called outcomes, and a ten fold difference in what you owe.

None of that makes the $9.99 rate bad value. A qualified inbound lead is worth far more than $9.99 to almost any US B2B company, so the unit economics usually work. It just needs to be modeled as a sales cost against pipeline, not folded into the support budget where a $0.99 assumption is doing the work. If you are running both, split the two motions in your forecast and price them separately.

Intercom seats are billed separately from Fin and list at $29 Essential, $85 Advanced and $132 Expert per seat per month. The useful break-even is straightforward: an Expert seat at $132 buys about 133 Fin outcomes. For most teams that is a light day of conversation volume, which is the real argument for the agent and not a marketing one.

The comparison buyers actually run is against Zendesk, and it is not symmetrical. Zendesk does not publish a per resolution price anywhere. Its pricing page describes the model, saying you "pay only for customer requests that were successfully resolved by the AI agent, without any escalation to a human agent", and then names no figure. What Zendesk publishes instead is a resolution allowance denominated in dollars: $2 per agent seat per month on Suite Team and on every Support plan, $5 on Suite Growth and Professional, $10 on Suite Enterprise and Enterprise Plus, capped at $5,000 a year across all plans, with no carryover of unused allowance.

Notice what that means. To convert your Zendesk allowance into a number of resolutions you need a rate Zendesk has not told you. Third party analyses converge on roughly $1.50 per resolution on a committed pack and $2.00 pay as you go, but that is reported, not published, and we are labeling it that way deliberately. Against Intercom's published $0.99, that would put Fin around 34% below the committed rate and about 51% below the pay as you go rate, on unit price alone.

All Zendesk figures above were read off zendesk.com/pricing and Zendesk's own resolution allowance documentation on September 5, 2026.

Here is the derivation that changes how the two rate cards compare, and it falls straight out of Zendesk's own documentation rather than out of anyone's opinion.

Zendesk grants the AI allowance per agent seat per month, but caps the total at $5,000 a year for every plan. Suite Enterprise grants $10 per seat per month, which is $120 per seat per year, so the cap binds at $5,000 divided by $120, or about 41.7 seats. Past roughly 42 Enterprise seats, hiring another agent adds a seat license and adds nothing to your included AI resolutions. On Suite Professional at $5 per seat per month, or $60 a year, the cap binds at about 83 seats. On the $2 plans it binds at about 208 seats.

The right hand column is the one to sit with. Even on the most generous plan, the included allowance covers roughly six or seven automated resolutions per agent seat per month at the widely reported committed rate, and about five at the pay as you go rate. A support team where each human handles a few hundred conversations a month will exhaust that in the first day or two and spend the rest of the month in overage. The included allowance is a sampler, not a budget, and any comparison that treats it as free capacity is wrong by an order of magnitude.

The last column depends on a rate Zendesk does not publish, so treat it as directional. The seat threshold in the third column does not: it comes only from Zendesk's published allowance and its published cap, and you can check the arithmetic yourself.

The framing of Fin versus Zendesk assumes a decision that many teams do not have to make. Fin runs natively on Intercom and also on Zendesk, Salesforce, HubSpot and Freshdesk, which means the AI agent and the helpdesk are separable purchases.

That opens a third option most comparison pages skip. Keep your Zendesk seats, your macros, your views and your reporting exactly where they are, and put Fin on top at $0.99 per outcome. You bypass Zendesk's allowance meter entirely and you buy the resolution layer at a published rate rather than a quoted one. At 1,000 automated resolutions in a month that is $990 through Fin against roughly $1,500 committed or $2,000 pay as you go through Zendesk, on top of Zendesk seats you are paying either way.

The honest counterweight: running two vendors means two contracts, two support relationships and two places where reporting lives, and Zendesk's native agent is already wired into the ticket data, the macros and the routing rules you have built. If the delta is a few hundred dollars a month, the integration tax is real and the native option is usually the calmer choice. If you are resolving tens of thousands of conversations, the unit price difference stops being a rounding error and the second contract pays for itself many times over.

Fin prices well when three things are true at once. Your conversation volume is high enough that human seats are the dominant cost, your questions are repetitive enough that a documented answer exists, and your help center is good enough that Fin has something to read. Under those conditions $0.99 against a $132 seat is not close.

It prices badly in the opposite corner. A low volume team with complex, bespoke conversations will see a poor resolution rate, pay the 50 outcome minimum, and get a bill for the handful of easy questions Fin caught while every hard one still reaches a human. Below roughly 1,000 inbound conversations a month, the arithmetic that justifies an AI agent is thin regardless of which vendor you pick, and the $49.50 floor is not the reason. The reason is that you have not yet accumulated enough repetition for automation to have anything to automate.

The middle case is the interesting one, and it is where the 14 day trial earns its keep. Intercom runs the trial with no cap on outcomes and no credit card, which is unusual and means you can measure your own resolution rate on your own tickets before you sign anything. Do that first. Every forecast on this page, including ours, is a placeholder for the number your own help center produces.

Intercom publishes a startup program offering 93% off Intercom and an entire year of Fin completely free. Read that against the rest of the page: a year of unlimited outcomes at $0.99 each has no ceiling on its value, and for a company handling 10,000 conversations a month at a 50% resolution rate it is worth roughly $59,400 over the year.

Qualification criteria change, so check the current terms directly rather than trusting this paragraph in six months. But if you are an early stage US company evaluating AI support agents on price, this program is very likely the cheapest credible option on the market, and it is not close enough for the per resolution comparison above to matter to you at all.

One caveat that applies to every discounted first year: the resolution rate you build during it becomes the volume you pay full price on afterwards. Model year two at list before you architect your whole support motion around the free year.