Introducing MCP: A new way to connect Fin and your customer data to your business tools

Delivering great AI customer service depends on more than just sophisticated AI models – it relies on real-time data and access to business systems like CRMs, billing platforms, and past conversations.

Without these connections, AI agents are unable to understand customer history and resolve questions as effectively.

The Model Context Protocol (MCP) provides a standard way for AI to securely connect to your business systems – without building custom integrations. With MCP, businesses can integrate AI faster, scale it across more systems, and deliver better support with less effort. In a world where speed and relevance define great support, MCP Connectors make it simple, secure, and scalable to unlock the value of AI across your tools.

MCP integrations rely on two key roles:

• Clients (like Fin) that request and use external data.
• Servers that expose data from business systems, like the Intercom Helpdesk.

At Intercom, we’re embracing MCP across both sides of the equation:

Fin as an MCP client

Fin can connect to MCP servers through a simpler, standardized setup. Non-technical users can quickly integrate Fin with a wide range of tools. Once connected, Fin can access data, perform actions in external systems, and more.

The Intercom MCP Server

Businesses can now make their Intercom conversation data accessible to external AI systems – uncovering insights, troubleshooting issues, and powering richer analytics with secure, scalable access.

Together, these updates make it easier for businesses to connect Fin to the tools they already use – and unlock more value from every customer conversation inside Intercom.

Connect Fin to your tools in minutes

Fin can already connect directly to your business tools, but soon this will become even easier. By acting as an MCP client, you can enable Fin to leverage our ready-to-use Intercom MCP Connector templates – or your own – to take actions in your business systems. For example, Fin can:

• Connect to the Stripe MCP server to look up a customer’s billing information.
• Connect Fin to real-time data from Shopify so customers can ask, search, or buy in natural language.
• Create or update support tickets via Linear’s MCP server.
• Access internal business data via custom-built MCP servers.

Easily configure and control MCP integrations

Within Intercom, these tools are exposed for Fin to use in “Data connectors” – flexible, no-code components that teammates can configure directly in the UI.

Integrating MCP with Fin gives your team full control over how external tools are used inside conversations. For each tool connected through an MCP server, teammates can:

• Decide when the tool should trigger, based on audience, channel, or conversation context.
• Choose which parameters Fin should collect, pre-fill defaults, or pull from Intercom data.
• Guide Fin’s behavior with clear instructions on how to use each tool.

Resolve complex queries with Fin Tasks

When using Fin Tasks to resolve complex queries like refunds, transaction disputes, or subscription issues, you can define how Fin uses MCP servers in these specific scenarios to give you greater control and flexibility. For example, you can:

• Trigger multiple actions in sequence to automate complex workflows.
• Set up conditions for when specific actions should be taken – like upgrading a user to a certain plan only when they meet the right eligibility criteria.
• Add safeguards for higher-risk actions, such as issuing refunds, ensuring Fin only acts when conditions are true.

This gives you the precision of rule-based automation with the power of AI – all directly configurable from your Intercom workspace.

Simplified setup, powerful results

The more connected Fin is to your systems, the more effective it becomes at understanding customer needs, solving problems, and delivering truly exceptional support experiences.

Getting set up is simple, as MCP servers handle the technical complexity, meaning non-technical teammates can connect Fin to tools without touching headers, payloads, or code. The same goes for connecting internal tools: MCP offers a simple, consistent way to do it, without any heavy engineering effort.

Make Intercom data accessible to your wider team

The new Intercom MCP Server, powered by Cloudflare’s infrastructure, lets external AI systems connect directly to your Intercom data – securely and in real time.

This means you can bring richer customer context into your existing AI tools, like Claude, used by teams across your business. Customer context, like conversations, user profiles, and tickets, are no longer limited to the Intercom Helpdesk, but are now available to help more teams piece together insights across conversations and user attributes.

With access to this data, teams can:

• Ask real-time questions about customer conversations to identify patterns needed for troubleshooting and debug issues in tools like Claude.
• Sync support data into engineering tools like Linear or issue-tracking platforms like Jira to enrich bug reports with real user context.
• Retrieve and interpret customer data from Intercom via your internal tools to drive business decisions across marketing, product management, and more.

This enables you to better understand your customers, optimize operations, and deliver better, more personalized customer experiences.

A new benchmark for AI-powered customer service

Open standards like MCP are reshaping how AI connects to everyday operations, making agents more powerful, flexible, and useful than ever before.

At Intercom, we’re proud to be at the forefront of this shift, enabling Fin to integrate more deeply into our customers’ systems and unlock valuable insights from their support data.

As the MCP ecosystem evolves, we’re excited to continue building smarter, more connected support experiences alongside our customers and the wider AI community.