Three products, one workflow, zero shared memory. That's the daily reality for power users juggling Anthropic's Chat interface, Cowork collaborative environment, and Claudee Code](/tools/claude-code/).
The pattern is familiar to anyone using these tools heavily: you sketch an idea in Chat, move to a Cowork session for structured back-and-forth, then hit a coding problem that pulls you into Claude Code. The work is continuous. The context evaporates at every tab switch. You re-explain details you already gave the AI three minutes ago, in a different product, from scratch.
Anthropics has acknowledged the gap. Team members have said in interviews that more overlap between the three products is coming, but no public timeline exists. The obvious fix is shared project memory across all three surfaces. If you've already briefed the assistant in Chat on a project - the goal, the constraints, the audience - that context should travel automatically into a Claude Code session. Anthropic's Projects feature already does something close to this within Chat, linking conversations to shared instructions and uploaded files. Extending that across product lines is the next logical step.
AI-assisted work rarely respects product category boundaries. A content strategy session turns into writing, which turns into building a small automation. Requiring users to manually re-establish context at each handoff is friction that undercuts the whole point of having a persistent AI assistant.
Until unified context ships, the practical workaround is a running context document you paste at the start of each new session. It works. It shouldn't be necessary.