Fin
Fin is Intercom's AI agent for customer service, resolving complex queries across voice, email, chat, and social with the Fin AI Engine and outcome-based pricing. Built for support teams running on Intercom, Salesforce, HubSpot, Zendesk, Freshdesk, and other major helpdesks.
Use Cases
Resolve complex support queries across chat, email, and voice
Train agents on Procedures, knowledge sources, and policies
Test agent behavior with simulated conversations before launch
Plug into existing Salesforce, HubSpot, Zendesk, or Freshdesk helpdesks
Analyze conversations with CX Score and AI Topics
Deploy across WhatsApp, SMS, Instagram, and Facebook Messenger
Pros
Outcome-based pricing means you pay only when Fin resolves a ticket
Procedures train the agent on policy and workflow specifics
Simulation testing previews agent behavior before launch
Fin AI Engine layers retrieval, reranker, and accuracy validation
Works on Intercom or any major external helpdesk
Cons
Minimum of 50 outcomes per month gates out smaller teams
Helpdesk seat cost separate from outcome charges on Intercom plan
Outcomes can rack up fast on high-volume support flows
Best-in-bake-off claims rely on customer-run tests, not third-party benchmarks
Voice channel restricted to enterprise contracts via sales
Platforms
Web
API
Chatbot
Compliance & Certifications
SOC 2
AICPA
ISO/IEC 27001
ISO/IEC
ISO/IEC 27018
ISO/IEC
ISO/IEC 27701
ISO/IEC
GDPR
European Union
CCPA / CPRA
State of California
Fin is what happens when Intercom rebuilt their support bot around the assumption that you'd pay only for resolved tickets, which is the right alignment but also why a small team will hit the 50-outcome monthly minimum before the bot has learned anything. Procedures let you train the agent on policy in plain language instead of decision trees, and that's the part Intercom keeps refining.
Must Try