Intercom Fin Pricing Explained, and What a Flat-Plan Alternative Costs

If you have searched for "intercom fin pricing" and come away unsure what your bill would be, you are not alone. Intercom publishes a headline number, $0.99 per Fin outcome, but the actual monthly cost depends on how many conversations you get, what fraction Intercom counts as resolved, which Intercom plan sits underneath, and how many seats you need on it. None of those appear on the headline.

This post is for founders, support leads and ops people who want a number they can put in a budget. It explains how Intercom Fin pricing works, what counts as a resolution, and what the underlying seats cost, then works a full example for a team with 2,000 conversations a month at three different resolution rates. It closes with the cases where Fin is still the sensible choice, because there are several.

All prices below were read from Intercom's pricing page and help center in September 2026 and re-checked on 23 October 2026. Pricing changes often; check their pricing page for current numbers before you decide anything.

How does Intercom Fin pricing work?

Fin is Intercom's AI support agent: it answers customer questions from your help content, runs configured procedures, and hands off to a person when it cannot help. Intercom prices it per outcome, a billing event that Intercom counts when Fin is judged to have finished a job in a conversation.

As of October 2026, the Intercom pricing page lists Fin at $0.99 per outcome, whether you buy it with an Intercom plan or run it on top of another help desk such as Zendesk or Salesforce. You are charged for at most one outcome per conversation, even if Fin answers several questions in it. Unused outcomes do not roll over.

There is also a floor when Fin runs on another help desk. Intercom's pricing page says a minimum monthly commitment applies to Fin with an existing help desk, and gives 50 outcomes as the example; the Fin standalone pricing page states a 50-outcome monthly minimum for Fin on your current help desk and shows none for Fin with Intercom. Fin's own help center describes this as a base plan of $49 a month that includes 50 resolutions, and says unused resolutions from that base are lost at the end of the month. Larger commitments are negotiated with sales rather than listed.

The pricing page also mentions a 14-day free trial that includes Fin, and an Early Stage startup program with steep discounts and a year of Fin free. If you qualify for that program, the arithmetic below changes for a year.

What counts as a Fin resolution (or outcome)?

This is the part that decides your bill, so it is worth reading Intercom's own wording. The Fin AI Agent outcomes article in Intercom's help center lists five kinds of billable outcome, at most one per conversation:

1. Resolution. No further help is requested after Fin's last answer. This has two flavours. A confirmed resolution is when the customer replies with something affirmative such as "Ok thanks" or "That helped". An assumed resolution is when the customer simply stops replying. Fin's help center article on resolutions puts the window at 24 hours: if the customer disengages for 24 hours after Fin's last answer, it counts.
2. Procedure handoff. Fin completes a procedure you configured that ends by handing the conversation to a human or a workflow. That handoff is billable because the procedure ran.
3. Qualification and disqualification. For teams using Fin on sales conversations, matching or not matching a prospect against your criteria is an outcome. Intercom lists qualification at a higher price, $9.99, and a disqualification at $0.99, so check the standalone pricing page if you plan to use Fin for lead capture.
4. Self-serve routing. Also a sales outcome: Fin matches a prospect to a routing outcome you marked as self-serve, billed at $0.99.

Intercom is equally clear about what does not count. If Fin only responds to a greeting, that is not an answer and no resolution is counted. Conversations that Fin escalates through its default behaviour or your workspace rules, without a procedure, are not billed. A procedure that fails is not billed. A clarifying question that the customer never answers is not billed either. And if a customer comes back to the same conversation for more help after a resolution was counted, even in a later billing period, Fin's help center says that resolution is deducted and not charged.

One more note. In mid-2026 Intercom changed how dashboards report Fin's resolution rate, excluding conversations where Fin was active but could not answer. The performance-metrics update says the definition of a resolution and the billing did not change. A rate that looks better on the dashboard is not the same as fewer billable outcomes.

What do the Intercom seats underneath Fin cost?

Fin does not run on its own unless you already have another help desk. For most buyers it sits on an Intercom plan, and Intercom plans are priced per seat. A seat is a paid login for a person on your team who replies to customers.

As of October 2026, Intercom's pricing page shows three plans. The table gives list prices on annual billing; month to month they are $39, $99 and $139. On 22 October 2026 the page also showed new customers Essential at $19 a seat on annual billing and $29 monthly, with the list prices struck through.

Lite seats are free limited logins for colleagues who need to look but not reply; only full seats can send messages. Fin is available on every plan at the same $0.99 per outcome. Intercom also lists optional add-ons such as Copilot, an AI assistant for human agents, at $29 per agent per month on annual billing.

So a small team's Intercom bill has a fixed part (seats) and a variable part (Fin), and the variable part can be the larger one.

Worked example: 2,000 conversations a month at 30%, 50% and 70%

Take a support team with three people who reply to customers and 2,000 inbound conversations a month: a small SaaS or a mid-sized online store. Assume every conversation goes through Fin first, and that the "resolution rate" in each row is the share of conversations that end in a billable outcome. Intercom's marketing on fin.ai claims an average resolution rate of 76% across its customers, so 70% is not an aggressive top row; 30% is a realistic first month with a thin knowledge base.

Prices are the October 2026 list prices above; new customers on the $19 Essential offer would pay $57 for three seats instead of $87. Check their pricing page for current numbers.

Three things stand out.

First, the Fin line dwarfs the seat line at this volume. At 70%, Fin alone is $1,386 a month, sixteen times the cost of three Essential seats. If you have been comparing Intercom to alternatives on seat price, you have been comparing the small number.

Second, the bill rises as the product gets better. Improving your knowledge base from a 30% to a 70% resolution rate more than doubles what you pay Fin, from $594 to $1,386. That is the intended design: Intercom's position is that you pay when Fin delivers value. It is still an unusual budget line, because most software gets cheaper per unit as you use it more, not more expensive in absolute terms.

Third, look at cost per conversation rather than per resolution. At 70%, Fin costs about $0.69 for every inbound conversation, resolved or not. At 30% it is $0.30 per conversation, but then a human is handling 1,400 conversations a month.

At 500 conversations a month rather than 2,000, divide the Fin column by four: roughly $150 to $350 plus seats. For very small volumes on another help desk, the 50-outcome minimum, not the rate, is what you pay.

What does a flat monthly plan cost for the same team?

The alternative model is a flat plan: a fixed monthly price for the AI support agent, with limits on usage rather than a meter on outcomes. A flat plan charges the same amount whether the agent resolved 30% or 70% of conversations, and whether a resolution was real or assumed.

PepoChat is priced this way. Every feature is on the free plan, which includes 200 AI replies a month and 5 knowledge sources; Starter ($39) includes 2,000 replies and 3 members, Growth ($149) 10,000 replies and 10 members, billed monthly or yearly at ten months' price, and replies past the allowance are billed at $10 per 1,000 rather than per resolution. Prices can change; the pricing page has the current figures, and you can start on the free plan without a card.

For the worked example, the comparison looks like this. At two to four replies each, the team's 2,000 conversations come to roughly 4,000 to 8,000 AI replies a month, far past the free plan's 200-reply cap (past it, new conversations go to the team inbox with a friendly message rather than failing). On Starter that is $39 plus $20 to $60 of overage, so $59 to $99 a month for the three members; Growth covers the whole range for a flat $149. Either way the plan price is the whole AI bill: the same at 30% as at 70%, with no per-seat charge for those three members and no minimum commitment.

That is not the same as saying a flat plan is always cheaper. It wins when volume is high relative to the price and the resolution rate is decent; it loses if you have a handful of conversations a month on a help desk you already run, where Fin's 50-outcome minimum would be your entire bill. Compute your own row of the table above and put it next to the flat price.

You are also buying different things. Intercom's plans include a full help desk with ticketing, email and multiple channels. PepoChat is an AI support agent with a web widget, a team inbox for human handoff, a knowledge base built from your site and documents, and integrations that let the agent act on Stripe, Shopify, WooCommerce and help desk tickets. It has no native Slack, WhatsApp or Messenger channels. If you need the whole suite, the standalone widget versus full suite question comes before price.

What incentives does per-resolution pricing create?

The phrase "you only pay when Fin resolves something" is doing a lot of work, so it is worth being precise about who decides what "resolves" means.

Assumed resolutions are counted, not verified. Under Intercom's definition, a customer who reads Fin's answer, finds it wrong, and gives up is indistinguishable from one who finds it right and leaves. Both stop replying; both become an assumed resolution after 24 hours; both cost $0.99. Fin does not have to be right to be paid, it has to be the last message. To Intercom's credit, greeting-only replies and unanswered clarifying questions are excluded, and a customer who comes back to the same conversation for more help gets the charge removed.

The vendor and the buyer measure success differently. Intercom's resolution-rate dashboard now excludes conversations where Fin could not answer, which makes the rate look better without changing the count that is billed. The number in your Fin report is not the number on your invoice, and you have to reconcile them yourself.

Improving the product costs you money. Write better help articles and the rate goes up, so the bill goes up. Some teams respond by rationing which conversations Fin sees, for instance turning it off for certain audiences to keep outcomes down. That is a strange incentive to build into a support tool.

Wrong answers have a second cost. A conversation that Fin "resolves" incorrectly often comes back as a second conversation, an email, or a churned customer. Intercom removes the charge only when the customer returns to the same conversation; otherwise you pay the $0.99, then you pay a person to fix it. That is the strongest argument for judging an AI support agent on the quality of its escalation rather than its price: a bot that says "I don't know" and hands off costs nothing under Fin's rules, and nothing in trust. Our guide to stopping chatbot hallucinations covers how grounding and citations keep the wrong-answer rate down whichever product you use.

None of this makes per-outcome pricing a bad model. It aligns the vendor's revenue with your automation rate, which is reasonable. It does not align with correctness, so budget on the assumption that some fraction of what you pay for was not actually resolved.

How to audit a Fin bill before and after you sign

Whether you are already on Fin or evaluating it, four checks take an afternoon and usually change the estimate.

1. Count conversations, not tickets. Fin bills per conversation with an outcome. Pull last quarter's monthly inbound conversation count from your current tool, including chats answered in one message. Those short ones are exactly what Fin resolves.
2. Estimate the assumed share. Sample fifty resolved conversations. How many ended with a "thanks" and how many simply went quiet? Both are billed under Fin's rules, but the quiet ones are where wrong answers hide.
3. Check where handoffs happen. Handoffs that come from Fin's default escalation are free; handoffs that come at the end of a procedure you built are billed. If you plan to use procedures for things like "collect the order number then route to a human", each of those is an outcome.
4. Set the alerts. Intercom says Fin has configurable alerts and limits. Set a monthly outcome cap at your budget before the first month, not after the first invoice.

If you want a benchmark for what resolution rate to expect by industry before you plug numbers in, our ticket-deflection benchmarks post gathers published figures; PepoChat has no deflection statistics of its own, and we do not invent them.

When is Fin still the right choice?

A balanced answer has to include the cases where the arithmetic above does not matter, or where Fin wins anyway.

You are already on Intercom. If your team lives in the Intercom inbox, has years of conversation history there, and pays for seats regardless, Fin is an incremental switch with no migration. The only new number is the outcome line, and you can trial it for 14 days with the outcome cap set low. Switching tools to save on the AI line while giving up the help desk your team knows is rarely worth it.

Your conversations are worth a lot. For a B2B product where a support conversation is tied to a five-figure account, $0.99 is noise. The question is whether Fin answers well, not what it costs. Per-outcome pricing is only painful when the conversation itself is low-value and high-volume.

You need many channels in one place. Fin's marketing lists chat, email, voice, Slack and social. PepoChat's widget is web only; if your customers write in on WhatsApp, Instagram and the phone, you need a multichannel help desk, and Fin is one of the more capable agents that runs inside one.

You have very high volume and a sales team. At tens of thousands of conversations a month you will negotiate a commitment with Intercom rather than pay list price, and qualification outcomes start to matter if Fin also fields sales chats. Check the limits on any flat plan before assuming it is built for that scale.

Your compliance needs are specific. Intercom's Expert plan lists SSO, HIPAA support and SLAs. PepoChat does not claim SOC 2, ISO or HIPAA certification; if your procurement checklist requires them today, that decides the question before pricing does. Our GDPR and chatbots guide lists what any buyer should ask of any vendor, including us.

What to do next

• To compare the tools side by side, read our roundup of the best Intercom alternatives for smaller teams. Work your own row of the table: last quarter's monthly conversations, times a resolution rate between 30% and 70%, times $0.99, plus seats. Put that next to the flat price on our pricing page. If the flat plan wins and your support is web-based, build a knowledge base from your site and documents, which the training guide walks through, then watch a week of real conversations in the team inbox. If Fin wins because you are already on Intercom or need channels we do not have, set the outcome cap first and audit the assumed resolutions after month one. Either way, the human handoff post explains the escalation behaviour that keeps wrong answers from becoming paid ones. If Zendesk is the other option on your list, Intercom vs Zendesk works the same four-person bill on both.

If you are weighing a switch rather than a top-up, the Intercom alternative page puts Fin's seats and outcomes next to PepoChat's flat plans at two volumes.