Your Kolinea folders in Claude

To talk to Claude about a folder, you used to copy what it holds into Claude, then copy its answer back into Kolinea. Once Claude is connected, it reads your folders itself: their threads, including your team’s internal notes, their documents, their participants, your directory and your templates.

It reads exactly what you read, no more and no less. It works in the workspace you point it to, and sees no folder you are not part of. When you ask it what to do first, it relies on the Cockpit: threatened milestones, then your actions in order.

What Claude prepares for a client waits in the folder’s planned publications, marked as prepared by your assistant. It gives you the publication’s link, which opens it ready to review. You correct it if needed, then publish it from Kolinea, or dismiss it.

Some actions only concern your team: an internal note, a participant added without being notified, a folder opened or duplicated. Claude does them straight away. A contact added to a folder only receives their link when you invite them, or when you publish something they can see.

The connection gives Claude your rights in Kolinea, no more. The audit log keeps what it did, as “Claire Martin (via her assistant)”, with your own name. You revoke it at any time from the “AI assistant” page, where it is listed with the date it was last used.

Kolinea adds no AI of its own. The content Claude reads only goes to Anthropic, the provider you chose: choose it as you would any tool you trust with your clients’ exchanges. What the assistant reads and prepares is detailed on the AI assistant page.