AI memory
We explore how AI can retain, understand, and use information over time, building a continuing understanding of people and their environments.

Research direction
Continuity in understanding can make help more meaningful over time. We research how AI can connect past conversations with present circumstances and update its understanding as lives and goals change, reducing the need to explain everything from the beginning. We also study how information from different stages can stay relevant to real life, and how people can easily add to or correct their circumstances and wishes, keeping long-term assistance flexible.
Applications we are exploring
Studying how ongoing learning and creative work can build on previous progress and current goals.
Exploring continuity in the understanding of project context and decisions as teams move between stages.
Let every conversation begin with understanding.
Useful assistance calls for an ongoing understanding of people and their lives. We explore how past conversations, present changes, and future goals can connect, so help can build from one interaction to the next.

- 01
The first conversation
A family would like to visit a park at the weekend. Assistance begins with understanding everyone's preferences, available time, and hopes for the outing.
- 02
Life changes
Plans and needs can change. Understanding people must evolve too, staying close to their lives as they are now.
- 03
Meeting again
The next time the family discusses an outing, assistance can build on what is already understood, reducing repetition while leaving room for new choices.
This is a future scenario we are exploring to illustrate our research direction, not a description of existing features.
Discuss this research area
We welcome conversations about business partnerships, technology collaboration, and joint research.




