Sales & Customer Support · Intercom MCP + Claude
Intercom MCP with Claude: Faster Support Replies and Cleaner Customer Records
The short answer
Claude works with Intercom through the Model Context Protocol (MCP), an open standard that acts as a secure adapter between the two. Once connected with an API key, Claude can read conversations, look up contacts and companies, fetch Help Center articles, and draft or post replies, within permissions your organisation sets.
What is Intercom MCP?
Intercom is a customer messaging platform used by sales, support and success teams to talk with customers in chat, email and in-app messages. It stores conversations, contact and company records, tickets and Help Center articles in one workspace. Intercom provides a hosted MCP server that exposes these resources through a stable interface, so AI agents and chatbots can work with them safely using natural language.
How Claude connects to Intercom MCP through MCP
The Model Context Protocol is an open standard that lets Claude securely use other software. Think of it as a universal adapter between Claude and a business tool. Intercom's hosted MCP server is the adapter that gives Claude access to your workspace, using an API key that your team creates and controls.
Once connected, Claude can read conversations, look up contacts and companies, search tickets and retrieve Help Center articles. It can also draft replies or, where you allow it, post them. Claude works through the same resources a support agent sees, so answers can be grounded in your existing knowledge base.
Permissions and approvals keep people in charge. Access is limited to the scopes your administrators grant, typically using a dedicated key with the least access needed. Consequential actions, such as sending a reply to a customer or changing a contact record, can require a human to review and approve before they happen. Every action is logged for audit.
What you can automate
Drafting replies to customer conversations
Claude reads the full conversation history and the relevant Help Center article, then drafts a reply in your brand voice. A support agent reviews the draft, edits it if needed and sends it. This shortens response times while keeping a person responsible for every message.
Summarising long customer threads
When a conversation runs to dozens of messages, Claude can produce a short summary of the issue, what has been tried and what the customer is waiting for. Account managers and new agents picking up a case can get up to speed in seconds rather than reading the whole history.
Keeping contact and company records tidy
Claude can review contact and company details against recent conversations and flag missing or outdated fields, such as job titles, company names or preferred languages. Staff approve suggested updates before they are saved, which helps sales and support work from accurate data.
Finding the right Help Center article
Support teams often know the answer exists but struggle to find it quickly. Claude can search your Help Center articles, match them to a customer's question and suggest the best link or excerpt to include in a reply. This makes new team members more effective and reduces repeat questions.
Routing and triaging tickets
Claude can read incoming tickets and suggest a category, priority and owning team based on the content and the customer's history. Your team sets the rules, and Claude's suggestions can be checked before tickets are reassigned, so urgent issues reach the right people sooner.
An end-to-end workflow: Escalated complaint handling with CRM follow-up
TriggerA new conversation from a customer containing words such as 'cancel', 'refund' or 'not working' arrives in Intercom.
Claude reads the conversation and the contact's history in Intercom and checks the company record.
Claude searches Help Center articles for matching troubleshooting or refund policy content.
Claude drafts a reply and a short case summary, then places the ticket in the correct team queue for review.
A support lead reviews and approves the reply, which is then sent from Intercom.
Claude updates the matching account record in your CRM (such as Salesforce or HubSpot) with the case summary, and notifies the account manager in Microsoft Teams.
The ticket is closed in Intercom once the customer confirms the issue is resolved.
OutcomeTypically, teams find that complaints reach a human decision faster, account managers see the full picture in their CRM, and customers receive consistent, policy-based replies. Results vary by team size and volume.
How we set it up
Book a free automation audit
We review how your sales and support teams use Intercom today, identify the conversations and records worth automating, and agree on the outcomes you want.
Set up secure access with an API key
We help your administrator create a dedicated Intercom API key with the least access the workflow needs. The key is stored in a secure vault, never in shared documents or chat messages.
Connect Claude through the MCP server
We connect Claude to Intercom's hosted MCP server, then test read and write permissions one by one so nothing is granted beyond what is required.
Design and test the workflow
We build the workflow across Intercom and your other systems, such as your CRM, Microsoft 365 or ticketing tool. We test it against real past conversations and refine the approval steps.
Launch, monitor and train
We go live with audit logs and monitoring in place, then train your agents, team leads and administrators on reviewing Claude's suggestions and handling exceptions.
Who benefits
- Support team leads
- Faster, more consistent replies with a clear review step before anything reaches a customer.
- Sales and account managers
- Up-to-date contact and company records, and summaries of customer history before every call.
- Customer success managers
- Early warnings from recurring issues in conversations, helping them act before a customer churns.
- IT and operations heads
- A governed, audited connection with scoped API keys and monitored access to customer data.
Security and control
- Use a dedicated API key with the least access needed, and rotate it on a regular schedule.
- Require human approval before Claude sends customer messages or changes contact records.
- Keep full audit logs of what Claude reads, drafts and changes in Intercom.
- Restrict which conversations and data sets Claude can access, based on team and role.
- Test thoroughly with real past cases before launch, and monitor performance after go-live.
For teams in the US, UK, India and the Middle East
Organisations in the United States and the United Kingdom can align the setup with state privacy rules and UK GDPR, with data handled under documented controls. In India, the setup considers the Digital Personal Data Protection Act, 2023, including consent and purpose limitation for customer data. For organisations in the UAE and Saudi Arabia, we design around local data protection laws, and we choose the hosting region for the Claude deployment to meet your residency requirements.
Learn more about Claude and MCP
Anthropic's AI model and assistant.
Secure links between Claude and business tools.
How engineers build and change software with Claude.
Hosted agents that run multi-step work.
The open standard Claude uses to connect to tools.
Frequently asked questions
What does the Intercom MCP server do?
Intercom's hosted MCP server gives AI tools a standard, stable way to reach your workspace resources, including conversations, contacts, companies, articles and tickets. It lets an AI assistant read this information and perform permitted actions, such as creating replies, using natural language requests instead of custom code for each task.
Is connecting Claude to Intercom secure?
Security depends on how it is set up. Access uses an API key, which we recommend scoping to the minimum permissions needed, storing in a secure vault and rotating regularly. Human approval steps, audit logs and monitoring add further control. We review these controls with your IT team before launch.
What can Claude do with Intercom once connected?
Typically, Claude can read conversations, look up contacts and companies, search tickets, fetch Help Center articles, summarise threads and draft or create replies. The exact actions depend on the permissions you grant. Sensitive actions can be restricted so that a person must approve them.
How long does setup take?
Many projects take a few weeks from the free automation audit to launch, depending on how many systems are involved and how much testing is needed. A single read-only assistant can be simpler than a full workflow spanning your CRM and ticketing tools. We agree on a timeline during the audit.
Which plan or permissions do I need?
Generally, you need an Intercom workspace where your administrator can create an API key with the access scopes your workflow requires. The specific plan features and permissions depend on your Intercom subscription and internal policies, so we confirm these during the audit rather than assume them.