Intercom (Fin AI Agent)

Intercom (Fin AI Agent) is worth it if your support team receives a high volume of repetitive queries and you need to automate a big chunk of that load without losing reply quality. Intercom was for years one of the most widely used customer support platforms, and today its differentiator is Fin, its AI agent that resolves support queries end to end without human intervention in a large share of cases.

Unlike a traditional chatbot with pre-built answers, Fin understands the context of the query, searches the company’s knowledge base and resolves or escalates as needed, learning from every interaction. Worth noting: the company is going through a brand restructuring (it rebranded corporately as “Fin”) and its acquisition by Salesforce was announced, something to keep in mind if you’re evaluating it for the long term.

In practice

This shows up in concrete workflows for a support team: a customer writes asking about an order’s status and Fin replies by checking the relevant system without escalating to a person, a more complex query gets automatically routed to a human agent with the whole conversation already summarized, or a lead gets its needs automatically qualified before a salesperson picks it up. For the human agents who do step in, the Copilot add-on works as an assistant that suggests replies and pulls relevant knowledge-base information while they work, without resolving the query autonomously the way Fin does. Fin also handles conversations consistently across email, web chat and other channels connected to Intercom, so a customer can start a conversation on one channel and continue it on another without losing context. For teams receiving hundreds or thousands of repetitive queries a month, automating that first layer frees up real time from the human team for the cases that actually need judgment.

Plans and pricing (2026)

• Essential (USD 39/agent/mo, USD 29/mo billed annually) — the entry tier for startups and small teams, with Fin available as a per-outcome add-on.
• Advanced (USD 99/agent/mo, USD 85/mo annual) — adds deeper automation and reporting tools for growing teams.
• Expert (USD 139/agent/mo, USD 132/mo annual) — built for large organizations, with security features and multibrand support.
• Fin AI Agent (USD 0.99 per outcome) — billed separately from the seat, on every plan; an “outcome” is a resolved query, a handoff with context, or a qualified lead.

There are also add-ons with their own cost (Copilot for human agents, conversation analytics, proactive support), and if Fin is used standalone, integrated into another support platform (Zendesk, Salesforce, HubSpot), there’s a monthly minimum equivalent to about 50 resolutions. Intercom also runs an early-stage startup program with discounts of up to 93% and a year of Fin free.

For this review we looked at the plans and pricing model published on Intercom’s official page, focusing on how the seat cost combines with Fin’s per-outcome charge, which is the hardest part of the total price to estimate in advance.

What to know before choosing

The most real limitation is exactly that: “pay per outcome” pricing can become unpredictable at high volume, since there’s no fixed monthly spending cap unless you explicitly set one up. On top of that, you have to pay the base Intercom seat regardless, so the total cost is never just Fin’s resolution charge — you need to add both parts to get the real number. The recent Salesforce acquisition also adds an open question about where the product is headed in the medium term.

Worth it for teams with high support volume where automating repetitive queries produces a clear, measurable saving. If your query volume is low or variable, calculate how many monthly “outcomes” you’d realistically generate before committing, because the cost can scale faster than expected. For small teams just starting to evaluate AI support, it’s worth checking the early-stage startup program before paying full list price.