Resolve Customer Tickets Automatically With Intercom's AI
Your support inbox is growing faster than your team. You're hiring more agents just to keep up, and response times are still slipping. Intercom's AI agent resolves common support questions instantly using your knowledge base — no human needed for the routine stuff.
Tools You'll Need
| Tool | What It Does | Cost | Link |
|---|---|---|---|
| Intercom | AI-first customer service platform where an AI agent resolves tickets from your knowledge base, with human handoff for complex issues | $0.99 per AI resolution (human agent plans separate) | Get it → |
The Walkthrough
Step 1: Set Up Intercom and Build Your Knowledge Base
What to do: Sign up at Intercom. Before activating the AI agent, build your knowledge base — articles covering your most common questions: shipping, returns, pricing, account issues, how-to guides. Use ChatGPT or Claude to draft articles from your existing FAQ content.
Use this prompt to write knowledge base articles: Ask Claude: “Write a help center article for [business name], a [business type], answering this customer question: [paste question]. Format: a clear title, 2-sentence intro, answer or steps under 150 words, and a closing line pointing to further help. Friendly and direct — not corporate.”
Why you’re doing it: Intercom’s AI agent can only answer questions it has knowledge base content for. The quality of your knowledge base directly determines the quality of AI responses. More articles = more tickets resolved automatically.
What to expect: 2–3 hours to write 15–20 knowledge base articles. Start with your top 20 most common support questions — these cover 80% of tickets at most businesses.
Step 2: Activate Fin — Intercom’s AI Agent
What to do: Enable Fin (Intercom’s AI) in your settings. Fin reads your entire knowledge base and starts answering customer questions immediately. It generates natural, conversational responses — not rigid template replies. When it can’t answer confidently, it hands off to a human agent.
Why you’re doing it: Fin resolves routine tickets in seconds, 24/7. Customers get instant answers. Your human agents focus on complex issues that actually need a human. The cost is $0.99 per successful resolution — far cheaper than a human agent handling the same ticket.
What to expect: 15 minutes to activate. Fin starts resolving tickets immediately from your knowledge base content.
Step 3: Install the Messenger on Your Website
What to do: Add Intercom’s Messenger widget to your website. Customers can chat with Fin directly, search your help center, and get instant answers without emailing support or waiting on hold.
Why you’re doing it: The best support ticket is the one that never gets created. When customers can self-serve through the Messenger, they get faster answers and your ticket queue stays manageable.
What to expect: 10 minutes to install. Copy-paste a code snippet into your website or use a platform integration (Shopify, WordPress, etc.).
Step 4: Set Up Human Handoff Rules
What to do: Configure when Fin should hand off to a human: when it can’t find an answer, when a customer asks for a human, when a conversation involves billing disputes, or when sentiment is negative. Route different issue types to different team members or departments.
Why you’re doing it: AI handles the routine; humans handle the exceptions. Clear handoff rules ensure frustrated customers reach a human quickly, while straightforward questions stay automated.
What to expect: 15 minutes. Test by sending in a few tricky questions to verify handoff works correctly.
Step 5: Monitor Resolution Rate and Expand Knowledge
What to do: Check Intercom’s AI analytics weekly. Track: resolution rate (% of tickets Fin resolves without a human), common unanswered questions (gaps in your knowledge base), and customer satisfaction scores. Write new articles for any recurring unanswered question.
Why you’re doing it: Every unanswered question you add to the knowledge base increases your AI resolution rate, which reduces cost per ticket and improves response times. This gets better over time with attention.
What to expect: 30 minutes per week. Most businesses see 30–50% AI resolution rates initially, climbing to 60–80% as the knowledge base matures.
Confidence Level
This workflow is Beta — Based on Best Available Knowledge. Intercom is a leading customer service platform. Fin, their AI agent, is widely reviewed and well-documented.
What to Do If It Doesn’t Work
- Low resolution rate: Your knowledge base is too thin. Write more articles covering the questions Fin can’t answer.
- Customers frustrated with AI: Make the “talk to a human” option easy to find. Don’t force customers through AI if they clearly want a person.
- Need more help? Intercom Support or email us at hello@thenewsbakery.com.