Anthropic Unifies Claude Memory Across Chat and Cowork Products
Anthropic has quietly unified its memory systems across its two main products: the Claude AI assistant and the team workspace tool, Cowork. The change means that information Claude learns about a user in one context—conversations, preferences, project details—becomes available in the other, effectively creating a persistent profile that carries across both interfaces.
The rollout is opt-out rather than opt-in, which means users who do not actively change their settings will have shared memory enabled automatically. Anthropic has provided controls within account settings for users who wish to disable the feature or selectively clear stored information. For teams using Cowork, this also raises questions about how individual user memories interact with shared team knowledge bases.
The move reflects a broader industry trend toward personalization through persistent memory in AI systems. Rather than treating each conversation as an isolated interaction, models that retain context across sessions can offer more relevant and efficient assistance. However, the approach also puts data retention practices front and center, requiring users to be intentional about what they allow the system to remember.
For organizations evaluating AI tools for workplace use, the announcement underscores the importance of reviewing default privacy settings and establishing internal policies around what information employees should share with AI assistants.