Intercom Fin is the AI customer-service agent built by Intercom, the customer messaging platform founded in San Francisco in 2011 by Eoghan McCabe, Des Traynor, Ciaran Lee, and David Barrett. In May 2026 Intercom rebranded the company itself to "Fin," and in June 2026 Salesforce announced an acquisition for approximately $3.6 billion — a signal of how central autonomous AI support has become to enterprise CX strategy.
The product launched as a GPT-4–powered chatbot in 2023, iterated to Fin 2 on Anthropic's Claude models, and shipped Fin 3 at Pioneer 2025. Fin 3 uses a multi-LLM approach with a proprietary Fin AI Engine that applies Retrieval-Augmented Generation (RAG) over a company's knowledge base, live CRM data, and third-party sources. Across all Intercom customers, Fin 3 claims a 67% average autonomous resolution rate and has processed more than 40 million conversations.
What distinguishes Fin from simpler chatbots is its ability to take actions — not just answer questions. It can look up order status in Shopify, update a Salesforce record, or create a Jira ticket, enabling agentic workflows rather than passive FAQ retrieval. Fin Voice and Fin Vision extend coverage to phone and image-based interactions. The pay-per-outcome pricing model ($0.99/resolved conversation) is a differentiator: you pay for demonstrated results. However, this adds up quickly at scale, and Intercom's own support quality has been a persistent complaint from buyers.
Key Benefits
- Outcome-based pricing: Charges only for successfully resolved conversations, aligning vendor incentives with customer results rather than raw volume.
- Agentic actions: Goes beyond retrieval to execute tasks in connected systems like Shopify, Salesforce, and Stripe without human intervention.
- Omnichannel reach: A single AI agent handles chat, email, voice, and image-based queries from one configuration layer.
- Enterprise compliance: Meets HIPAA, SOC 2 Type II, GDPR, and ISO 42001 (AI management), making it viable for regulated industries.
Use Cases
- E-commerce deflection — Automatically resolves order status, return requests, and shipping queries by pulling live data from Shopify or Stripe, handling routine tickets without a human agent.
- SaaS tier-1 support — Answers product how-to questions grounded in help center content, routes edge cases to the right human team, and logs outcomes in Salesforce or HubSpot.
- Multilingual global support — Serves customers across 45 languages with consistent brand tone, reducing the cost of staffing multilingual agent teams.
- After-hours coverage — Maintains 24/7 resolution capability for common queries, escalating to human agents during business hours only when Fin cannot resolve the issue autonomously.
