Intercom Fin

Intercom Fin is an AI support agent that handles tier-1 customer questions inside Intercom's chat widget.

Point Fin at your help docs, structured FAQs, and past Intercom conversations. It ingests the content, builds a model of your product and policies, and starts answering customer questions in chat. When confidence drops below a configurable threshold, Fin routes the conversation to a human agent with full context preserved, so the customer doesn't have to repeat themselves.

The pitch is simple. Deflect tier-1 questions automatically and free up human agents for the work that actually needs them.

The pricing model is unusual and worth flagging. You pay per successfully answered question rather than per seat or per message. The vendor only makes money when Fin actually helps a customer, which aligns incentives between Intercom and your support team. There's no gap where you're paying for capacity you don't use.

Common workflows by role:

• Support leads configure Fin against the existing knowledge base
• Customer success teams monitor conversations and update content where Fin struggles
• Operations track resolution rate as a business metric tied to staffing decisions
• Engineers integrate Fin's API for in-product help

A few specifics about setup:

• Fin reads existing Intercom help center articles automatically
• Custom answers can be authored for high-frequency or policy-sensitive topics
• Confidence thresholds are tunable per topic
• Routing rules send escalations to the right team or queue

The product is tightly integrated with Intercom's chat widget, customer data, and conversation history, which makes it most valuable for teams already on the Intercom platform. For teams not on Intercom, it exists as a standalone offering, but the integration value drops significantly.

Setup involves:

• Connecting Fin to your existing Intercom workspace
• Pointing it at help docs and past conversations
• Configuring confidence and routing rules
• Reviewing answers and refining content where Fin underperforms

Fin's quality depends heavily on the quality of your documentation. If help docs are stale, contradictory, or missing key topics, Fin will hand off most questions to humans anyway. Teams with current, thorough documentation see the biggest deflection lift.

Per-resolution pricing model. Billed only when Fin successfully resolves a customer question rather than per seat or per month.

Best for support teams drowning in repetitive tier-1 questions where documentation is current and thorough. Not the right fit for specialized troubleshooting workflows or teams not on Intercom, where other AI support platforms may fit better.