Chatz

AI

Data connectors (MCP)

Let the bot look up live systems while it replies.

What a connector does

A data connector is different from the rest of the knowledge base: instead of teaching your AI chatbot a fixed answer, it lets the bot reach out to another system and read a live answer at the moment it replies. Chatz connects to these external systems through MCP, a standard way for an AI to call another service's tools.

  • A document or a saved entry answers from something written down in advance.
  • A connector answers from a system that changes constantly, such as a booking system, so the bot's answer is never stale.

Live lookups mid-reply

When a bot with a connector attached needs information it does not already have, it calls out to the connected system in the middle of composing its reply, then folds what it finds into the answer. A booking system connector, for example, lets the bot check real availability before telling a customer whether a date is open, rather than guessing from anything it was told earlier.

Scope what a connector can see
A connector only exposes what the connected system's own access allows. Set up and review connectors with the same care you would give any integration that can read your live business data.

OAuth consent for external clients

Chatz can also be the system being connected to. When an external AI client, such as Claude, wants to read your Chatz data through this same mechanism, you see a standard OAuth consent screen naming the client and what it is asking to access, and nothing is shared until you approve it.

That consent step is what keeps the connection two-way and safe in both directions: a bot inside Chatz reaching out to another system, and an outside AI client asking to reach into Chatz, both go through an explicit approval rather than a standing, invisible connection.