Anthropic has introduced a memory feature for its Claude AI model, currently available to users with Team and Enterprise plans, following the recent announcement of Claude’s ability to generate spreadsheets and presentations.
According to a company blog post, Claude’s new memory allows it to “remember you and your team’s projects and preferences.” This personalized experience is designed to improve efficiency and collaboration for professional users. The memory functionality is not available to free users or those on individual Pro or Max plans, highlighting a distinction in service based on the user’s subscription tier.
In addition to the memory feature, Anthropic has launched incognito chats for all users. Similar to the incognito mode found in web browsers, these chats are not saved in the conversation history or stored in Claude’s memory, thereby providing an added layer of privacy for users who require it.
For Team and Enterprise users, Claude can now retain professional context, including client needs, project details, and team processes. When the “projects” feature is utilized, the AI model can create separate memories for individual projects, offering a dedicated workspace for each endeavor and potentially enhancing organizational capabilities.
Anthropic acknowledges that the introduction of the memory function brings “new safety considerations.” The company emphasizes that the feature is designed for work-related tasks and is structured to avoid archiving sensitive information. While specific details regarding these safety considerations were not disclosed, Anthropic plans to evaluate and test the feature before considering its expansion to a wider user base.
Team and Enterprise users can enable Claude’s memory within the app’s settings. Here, they can also review and adjust memory summaries to suit their needs. Additionally, Enterprise administrators have the flexibility to disable the memory feature for all users under their plan, providing control over the implementation of this new capability.
Moreover, users have the ability to import memory from other AI programs into Claude or export their Claude memory. This interoperability positions Anthropic closer to its competitors, such as Google Gemini and ChatGPT, which have already implemented memory features and continue to expand upon them.
