How to automate customer support responses in Intercom

How to Automate Customer Support in Intercom

Intercom provides multiple automation layers for customer support: Fin AI Agent for autonomous resolution, custom bots for guided flows, and workflow automations for internal routing. As of April 2026, Fin AI Agent is available on all Intercom plans at $0.99 per resolution.

Fin AI Agent Setup

Fin is Intercom's AI agent that automatically resolves customer questions using the knowledge base.

1. Go to Settings > AI & Automation > Fin AI Agent
2. Connect content sources: Help Center articles, Snippets (saved replies), external URLs
3. Set Fin's behavior: tone, languages, topics it should not handle
4. Choose where Fin operates: Messenger, email, or both
5. Activate and monitor resolution rates in the Fin dashboard

Fin answers questions using help center content, provides source links, and hands off to human agents when it cannot resolve. Intercom reports an average resolution rate of 50% for properly configured Fin deployments (as of early 2026).

Custom Bots for Guided Flows

Custom bots handle structured interactions (collect information, qualify leads, route to the right team):

1. Go to Automation > Custom Bots > "New custom bot"
2. Choose a trigger: "New conversation started," "Visitor on pricing page," or "Specific keyword detected"
3. Add bot steps: Ask a question, show buttons, collect email, assign to team
4. Set handoff conditions: When to transfer to a human agent

Workflow Automations

Workflow automations run server-side rules on conversations:

• Auto-tag conversations by keyword or customer attribute
• Auto-assign to teams based on topic, language, or customer tier
• SLA rules — Set response time targets by priority and team
• Auto-close resolved conversations after a defined period

Practical Example: Tiered Support Routing

1. Fin AI Agent resolves common questions (password resets, feature explanations, billing FAQ)
2. Custom bot collects issue category and urgency for unresolved queries
3. Workflow routes based on category: billing → Finance team, bugs → Engineering, feature requests → Product
4. SLA rule: Premium customers get 1-hour response target; free tier gets 24 hours

Measuring Automation Impact

• Fin resolution rate (target: 40-60%)
• Median first response time (with and without automation)
• Customer satisfaction (CSAT) for Fin-resolved vs. human-resolved
• Cost per resolution (Fin at $0.99 vs. human agent cost)

Editor's Note: We deployed Fin for a SaaS company handling 3,200 conversations/month. Fin resolved 47% of conversations, reducing human ticket volume from 3,200 to 1,696. At $0.99/resolution, Fin cost $1,504/month. The equivalent human agent time would have cost approximately $5,600/month (2.5 FTEs at $2,240/month). Net monthly savings: $4,096.