Giving your AI more capabilities
Connect your AI product to live tools and data with an MCP connection, send results anywhere with webhooks, and let end users bring their own files.
Your AI Product is capable on its own — it runs the conversation, reasons over the answers and writes the result. These are the ways to extend it beyond that: live access to your own tools, a route for the data to travel, and a way for end users to bring their own material.
Connect any tool with MCP
The MCP node is the main way to give your product capabilities Productised doesn't build itself.
MCP (Model Context Protocol) is an open standard. Any service that publishes an MCP server can be connected — your CRM, your calendar, your database, a search provider, your own internal API. You supply the server URL and a token, choose which of its tools your product may use, and Claude can call them during a conversation.
MCP connections
How to add an MCP node, connect a server, and choose which tools your product can use.
This replaced a set of narrower built-in nodes. Rather than Productised picking a search vendor and a document store on your behalf, you connect the providers you already use, with your own account and your own credentials — so your data stays in your relationships, and the capability isn't limited to what we happened to integrate.
Send results wherever you need them
A webhook posts each completed response to any URL you choose, so leads and AI output can flow into your CRM, your email platform, a spreadsheet or an automation tool.
Webhooks
Post every completed response to a URL of your choice.
Zapier
Route responses into thousands of apps without writing code.
Let end users bring their own content
If your product needs to read something the end user has — a CV, a brief, a website — add an Upload or URL question in the AI Product node's Questions panel. The end user attaches it mid-conversation and Claude reviews it as part of the session.
This is the end user's content, specific to their session. It is not your reference library, and it is not retained as one.
Getting the AI to know your business
Everything the AI knows about your business comes from the prompt in the AI Product node's Collection Instructions. That is where your persona, your methodology and the way you talk about your work belong.
The Reference Material field in the AI Product node is not currently active. Anything entered there is saved but is not yet passed to the AI. Put context you need the AI to use in Collection Instructions instead.