It’s how an agent reaches Gmail, Slack, Google Drive, QuickBooks, or another system instead of staying trapped in chat. The access level is the real story, so always check it.
Think about giving an employee a keycard for one office suite. It doesn’t make them the owner of the building; it lets them into the rooms they’re allowed to enter. Some keycards open one room, some a whole floor, some only during business hours. A connector is that kind of access for an AI tool: it may be read-only, it may ask before every action, or it may be broad enough that you treat it carefully.
How it shows up
Connectors are what let AI pull context from a tool or act inside it. “Read my emails and flag the ones needing a response” is a connector-style task, so is “draft this reply in Outlook.” They show up as tools the agent can use, some built through MCP, some using an API underneath, some packaged by the product itself. The operator’s questions stay the same: what app does this open, what can it read, what can it write, and does it need approval before it acts? That’s why connectors are part of integration, wiring one working system to another, where the quality of the connection sets the quality of the work.
Why you care
Without connectors, an agent works only with what you paste in. With them, it works closer to the real source of truth. That’s powerful, but access is never neutral: AI gets more useful and more sensitive the moment it can touch your real tools. A connector is that moment.