WorkDesk is the desk you work at, rebuilt for AI-assisted work. It’s where the operator keeps the material that makes AI useful, especially the context that should survive beyond one conversation.
The founding analogy is a real desk. Fifty years ago, offices had to change because computers changed the work, and the old desk wasn’t built for screens, keyboards, and files moving digitally. Now AI is changing the work again, so the desk has to change again.
How it shows up
In practice, WorkDesk looks boring from the outside. It’s files and folders: a transcript lands as a source, a meeting note points back to the raw transcript, a rule tells the agent what it can and can’t do. That boring structure is the point. When Claude Code, Codex, or another agent opens the WorkDesk, it sees the same desk you see and can read the files, follow the rules, and work in a vault instead of remembering everything from one chat. The desk also keeps the operator in control, because if a tool changes or a model gets expensive, you still have the files and can switch. This is where WorkDesk connects to work architecture: work architecture designs the work, and WorkDesk gives that design a place to live and run.
Why you care
AI gets much more useful when it has a stable desk instead of a pile of disconnected conversations. The files become memory, the rules become guardrails, the sources become evidence, and the workflows become repeatable. The real asset is the working context you own, not the chat.