Intercom MCP
Access Intercom conversations, contacts, companies, Help Center articles, tickets, and workspace data through Intercom's hosted MCP server.
Verdict
Common use cases
- Triage overnight support tickets before standup
- Draft replies using full conversation history
- Pull customer feedback for roadmap planning
- Identify recurring issues across recent tickets
- Surface account context during sales calls
Integration
- Vendor
- Intercom MCP
- Category
- communication
- Auth
- API_KEY
- Composio slug
intercom_mcp
Tools
Setup
Setup guide
- 11. In Switchy, open your workspace settings and navigate to the Integrations tab. 2. Click 'Add Integration' and select Intercom MCP from the communication category. 3. In a separate tab, log into your Intercom workspace and go to Settings > Developers > Developer Hub. 4. Create a new app or use an existing one, then generate an API access token with read permissions for conversations, contacts, and tickets. 5. Copy the token and paste it into the API key field in Switchy, then click 'Connect'. 6. Switchy will verify the connection by fetching a test conversation—you'll see a green checkmark when it succeeds. 7. Open any Project, type '@Intercom' in the chat, and ask it to summarize recent support conversations to confirm the integration works.
What teammates see: by default, memories from Intercom MCP are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Projects have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
Triage Urgent Tickets
@Intercom pull all conversations from the last 24 hours marked as priority or unassigned, then summarize the top three issues that need immediate response.
Draft Context-Aware Reply
@Intercom get the full conversation thread for ticket #12847, then draft a reply that addresses their billing question and acknowledges the issue they reported last week.
Extract Feedback Themes
@Intercom review all closed conversations from the past two weeks and list the top five feature requests or complaints that came up more than once.
Customer Context Lookup
@Intercom find the contact record for sarah@acmecorp.com and summarize their last three support conversations, including any open issues.
Weekly Support Summary
@Intercom analyze all conversations closed this week, then create a summary showing total volume, average response time, and the three most common topics.
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@Intercom summarize the most common issues from conversations tagged 'mobile-app' this month and suggest which ones we should prioritize
Analyzed 67 conversations tagged 'mobile-app' in January. Top issues: (1) Login failures on iOS 17.2+ (18 reports, affecting paid users), (2) Push notification delays (12 reports, mostly free tier), (3) Offline mode sync errors (9 reports, enterprise accounts). Recommendation: Prioritize the iOS login issue—it's blocking revenue-generating users and has the highest urgency score (avg 8.4/10). The sync errors affect fewer users but hit your largest contracts.
This pairs Intercom data with AI reasoning to surface patterns. The AI interprets conversation content, tags, and user segments to generate insights. Accuracy depends on how consistently your team uses tags and conversation metadata. Best for weekly planning or quarterly roadmap reviews, not real-time triage.
Use-case deep-dives
When Intercom MCP helps a 6-person team triage launch-week tickets
A 6-person SaaS team launches a new feature and expects 200+ support conversations in the first 72 hours. The Intercom MCP lets the product manager and two engineers jump into the shared workspace, pull conversation threads by tag or user segment, and draft replies without leaving the AI context. The win is speed: instead of Slack screenshots and copy-paste, the team reviews actual conversation objects and writes responses that reference prior messages. The threshold: if your support volume stays under 50 conversations a week, the MCP is overkill—just use Intercom's native UI. Above that, especially during launches or incidents, the MCP turns your AI workspace into a real-time triage room. Worth it if your team already lives in a shared AI tool and needs to move fast on customer context.
When this MCP wins for quarterly feedback analysis
A 3-person growth team at a B2B startup runs quarterly user research by pulling Intercom conversations from their enterprise and SMB segments. The Intercom MCP lets them query conversations by custom attributes (plan tier, signup date, feature usage), then feed those threads into the AI workspace for thematic coding. The advantage over manual export: the MCP keeps conversation metadata intact, so the team can cross-reference user IDs, tags, and timestamps without building a spreadsheet. The limit: if your research questions need more than conversation text—like session replays or product analytics—you'll still need to stitch data from other tools. Best fit for teams that treat Intercom as their primary qualitative data source and want to analyze 100+ conversations without leaving the AI context.
When the MCP helps a solo founder spot missing help docs
A solo founder running a 200-user SaaS product uses Intercom for onboarding chat. Every week, they pull the last 50 conversations tagged 'onboarding' into the AI workspace and ask it to flag repeated questions that aren't covered in the help center. The Intercom MCP makes this a 10-minute weekly ritual instead of a manual review. The trade-off: if your onboarding flow is stable and you're not iterating on docs, the MCP won't surface much new signal. It's most useful in the first 6-12 months post-launch, when you're still learning what users don't understand. After that, the value drops unless you're shipping new features frequently. Right call for founders who want to close content gaps fast without hiring a support lead.
Frequently asked
What does the Intercom MCP do in Switchy?
It connects your Intercom workspace to Switchy's AI agents so they can read conversations, retrieve customer data, and answer questions about support tickets without switching apps. Your team gets instant context on user issues, conversation history, and customer segments directly in chat threads. The MCP uses Intercom's API under the hood, so it sees whatever your API key has access to.
Do I need admin access to connect Intercom?
You need an Intercom API key with read permissions at minimum. Most teams generate a dedicated key from Intercom's developer settings rather than using a personal token. If you want agents to update conversations or tag users, the key needs write scopes too. Check your Intercom workspace permissions before generating the key — standard agents can create keys, but admins control what those keys can access.
Can the Intercom MCP send messages to customers?
Not yet. The current integration focuses on reading data — pulling conversation threads, user attributes, and ticket metadata so your team can make informed decisions. If you need to reply to customers, you still do that in Intercom directly. We prioritise read-only access first because it covers 80% of support team workflows without the compliance risk of automated outbound messages.
How is this different from just opening Intercom in another tab?
Switchy's AI can query across multiple conversations, summarise trends, and pull customer context into your current work without manual searching. Instead of tabbing between Intercom and your project docs, you ask the agent 'what did this user report last week' and get an answer inline. It's faster for triage and handoffs, especially when you're juggling five support threads at once.
Who on the team should connect the Intercom MCP?
Whoever manages your Intercom workspace or has access to generate API keys. Once connected, everyone in your Switchy workspace can query Intercom data through agents — they don't need individual Intercom logins. If your support team uses Intercom daily but your engineers only need occasional access, this setup saves you seat licenses while keeping data accessible to the whole crew.