Intercom's Fin charges $0.99 per AI resolution on top of seats. Chatsy is flat. Honest comparison with real cost math at 500, 1,500, and 5,000 resolutions.
TL;DR:
- Fin is Intercom's AI agent, now branded as a separate product at fin.ai. Pricing is $0.99 per AI resolution on top of Intercom seat fees ($29 to $132 per seat per month).
- Fin is mature, well-integrated with Intercom, and tested across 25,000+ businesses per Intercom's public claim.
- The cost model punishes growth. At 5,000 monthly resolutions, Fin charges $4,950 in AI fees alone. Same volume on Chatsy Pro is $475 flat.
- Fin requires Intercom. It is not a standalone platform for teams that do not use Intercom for tickets.
- If you are already deep in the Intercom ecosystem and want a turnkey AI layer, Fin is fine. If you want predictable pricing, model choice, or standalone deployment, Chatsy fits better.
Fin works. That is the short version. The longer version is that Fin's per-resolution pricing creates a structural problem: the better the AI gets at resolving tickets, the more you pay. For some teams that math is fine. For many, by year two, it is not.
This article compares Fin to Chatsy on pricing, fit, and the things that actually break in production.
Fin started as Intercom's in-product AI agent in 2023, branded "Fin AI" inside the Intercom app. In late 2024 Intercom started pushing Fin as a separate product at fin.ai with its own marketing site, although the product still requires an Intercom workspace to deploy. The strategic intent appears to be positioning Fin as a standalone brand to compete with Decagon, Sierra, and Ada, while still tying it to Intercom's seat-based business model.
Fin is good at three things:
Mature out-of-the-box behavior. Intercom has run Fin against a large customer base for two years. The default behaviors, escalation paths, and tone calibration are battle-tested. You do not have to think hard about prompt engineering for it to work reasonably.
Native Intercom integration. If you already run Intercom for tickets, Fin reads your help center, your saved replies, your conversation history, and your customer attributes without any extra wiring. The integration is the deepest of any AI agent on the market.
Good defaults for routing and handoff. When Fin cannot answer, it hands off to a human agent inside the Intercom Inbox with full context. The handoff UX is more polished than most third-party AI tools because it was designed inside Intercom.
That is the fair summary. Now the structural problems.
Per Intercom's public pricing page, Fin charges $0.99 per AI resolution. A resolution is when Fin answers a question or completes a workflow without escalating to a human. Per Intercom's docs, escalations triggered by Fin's default fallback are not billed, so you only pay when it works.
This sounds reasonable until you do the math at scale.
| Monthly resolutions | Fin AI fees | Plus Intercom seats (5 seats Advanced) | Total |
|---|---|---|---|
| 500 | $495 | $425 | $920 |
| 1,500 | $1,485 | $425 | $1,910 |
| 5,000 | $4,950 | $425 | $5,375 |
| 10,000 | $9,900 | $425 | $10,325 |
A 5-seat Advanced plan at $85 per seat per month is $425 in seat fees. Add Fin resolutions on top.
Compare to Chatsy Pro at $475 per month for 15,000 credits and 5 seats. Assuming an average of 1.5 credits per resolution (a reasonable estimate for a mix of GPT-5.2 and Claude Opus 4.6 responses), 15,000 credits is roughly 10,000 resolutions per month.
| Monthly resolutions | Chatsy plan | Total |
|---|---|---|
| 500 | Hobby $35 | $35 |
| 1,500 | Standard $140 | $140 |
| 5,000 | Pro $475 | $475 |
| 10,000 | Pro $475 (at edge of credits) | $475 |
The pattern is clear. Fin scales linearly with success. Chatsy is flat within tier. A team whose AI handles 5,000 resolutions per month pays roughly $5,375 on Fin and $475 on Chatsy. The gap is 11x at that volume.
Enterprise customers get volume discounts on Fin (Intercom does not publish the curve), so the real number for high-volume customers is somewhere below $0.99. But the structural shape stays the same: every successful resolution costs money.
Fin is marketed at fin.ai as a standalone AI agent. In practice, it is not.
To deploy Fin, you need an Intercom workspace. To handle ticket workflows triggered by Fin, you need Intercom's Inbox. To use Fin's deep integrations with help center articles, you need those articles inside Intercom Articles.
This matters in two cases.
You do not use Intercom today. Adopting Fin means adopting Intercom. That is a much bigger decision than picking an AI vendor. Intercom seats, Intercom workflows, Intercom data residency, Intercom contracts.
You are leaving Intercom. Fin does not come with you. There is no Fin-on-Zendesk or Fin-on-HubSpot product. If you migrate off Intercom, you are also migrating off Fin and need to pick a new AI vendor. The migration playbook for leaving Intercom covers the mechanics.
Chatsy is the opposite. It runs as a standalone widget on your website regardless of what helpdesk you use. You can run it alongside Zendesk, Help Scout, Front, Intercom, or no helpdesk at all.
Intercom does not publicly disclose which foundation model powers Fin, and the answer is "it depends." Public statements from Intercom engineering suggest Fin runs a multi-model stack with OpenAI, Anthropic, and proprietary fine-tuned models depending on the task. The selling point is that you do not have to think about it. The drawback is that you cannot change it.
Chatsy exposes model selection. GPT-5.2, Claude Opus 4.6, Gemini 3 Pro, o3, DeepSeek V3.2, Llama 3.3 70B, plus roughly 25 others. Different models cost different amounts (1 to 5 credits per response). Some are cheaper for simple FAQs. Others are stronger on multi-step reasoning. The tradeoff is yours to make.
This matters more than it used to. Foundation models are improving in different directions: Claude is stronger on tool calling and instruction following, GPT is stronger on raw reasoning, Gemini is stronger on long context. Locking yourself to a vendor's choice in 2026 means you cannot benefit from improvements in any one model line.
| Dimension | Fin (Intercom) | Chatsy |
|---|---|---|
| Pricing model | $0.99 per AI resolution plus Intercom seats | Flat monthly tiers, credit-based usage |
| Standalone use | No, requires Intercom workspace | Yes, runs standalone on any site |
| AI model choice | No, proprietary multi-model stack | Yes, 25+ models including GPT-5.2, Claude Opus 4.6, Gemini 3 Pro |
| Setup time | Requires Intercom setup, then Fin enable | Under 1 hour, self-serve |
| Self-serve signup | Yes, but bundled with Intercom subscription | Yes, free plan available |
| Tool calling and workflows | Mature, built into Intercom | Available, configured in dashboard |
| Help center integration | Native (Intercom Articles) | Crawl any URL, upload docs, sync with most help centers |
| Conversation handoff | Polished, in Intercom Inbox | Built-in mailbox and live chat handoff |
| Best for company size | Existing Intercom customers | 5 to 500 employees, any stack |
| Public customer claim | 25,000+ businesses (Intercom claim) | SMB SaaS, ecommerce, agencies |
| Founded | 2023 (as Intercom feature) | 2023 |
These numbers assume a 5-seat support team on Intercom Advanced ($85/seat/month annual) for the Fin comparison, and Chatsy plans matched to the credit budget.
This is a small team early in AI deployment.
Caveat: at 500 resolutions, the average team is also doing 500 to 1,500 human-handled tickets per month, and Intercom's seat fees cover that. Chatsy's mailbox and live chat are included in the same plan but do not replace a full helpdesk if you have complex ticket workflows. Honest comparison: Fin is bundled with a full helpdesk. Chatsy is bundled with a lighter mailbox.
Mid-size SMB with active AI deployment.
Growing team where AI handles a real share of volume.
The gap widens as volume grows. This is the structural problem with per-resolution pricing.
Be honest about whether you fit this profile:
If three or more apply, stay on Fin. It is a fine product for that buyer.
Most teams reading this article:
If you are already on Intercom and Fin and you want to move, do not cancel Intercom on cutover day. Run both in parallel for two to four weeks while you validate the new platform. The migration playbook for leaving Intercom covers exports, what does not port (custom bots, Workflows, Series), and the rollback plan.
The main gotcha: Fin's behavior is tightly coupled to Intercom Articles. If you have built a lot of content inside Intercom's editor, plan time to export articles via the Articles API and re-host images that live on Intercom CDN URLs.
Honest disclaimers:
The most common Fin alternatives in 2026 are Chatsy (transparent flat pricing, standalone), Chatbase (popular self-serve AI chatbot), Eesel AI (help-center-focused), Ada (enterprise AI agent platform), Decagon and Sierra (enterprise with custom training), and Maven AGI (mid-market). On the helpdesk side, Zendesk's AI agents and Help Scout's AI Assist compete on bundled deployment.
No. Fin is marketed at fin.ai as a separate product, but the product still requires an Intercom workspace to deploy. To use Fin you need Intercom Articles for content, Intercom Inbox for ticket handoff, and Intercom seats for human agents. Adopting Fin without Intercom is not a supported configuration.
Fin's published price is $0.99 per AI resolution on top of Intercom seat fees. Seat fees range from $29 (Essential) to $132 (Expert) per seat per month on annual billing. A team with 5 seats and 1,500 monthly resolutions pays roughly $1,910 per month. Enterprise contracts get volume discounts on the per-resolution rate, but Intercom does not publish the discount curve.
It depends on your existing stack. If you already run Intercom for tickets, enabling Fin is the path of least resistance and the price at moderate volume is reasonable. If you do not run Intercom, adopting it just to get Fin is a much bigger commitment than picking a standalone AI vendor. At 50 employees with 1,000 to 3,000 monthly resolutions, Chatsy at $140 to $475 per month is meaningfully cheaper than Fin at $1,400 to $3,400 per month plus seats.
Fin is a mature product. If you are already deep in Intercom and your monthly resolution volume is not large enough to make the per-resolution pricing painful, stay on Fin.
If you are not on Intercom, or you are looking at the Fin bill and watching it scale faster than your support team, start with Chatsy. Free plan to test, transparent tiers on the pricing page, model choice, and no per-resolution fees. You can be live this afternoon.
Sierra AI is genuinely the best enterprise AI agent. For Fortune 500 with seven-figure budgets. For SMB and mid-market, Chatsy is the saner pick.
