Memory
Memory is how an AI product retains user preferences, facts, or past context across sessions, beyond the single context window.
Memory can feel magical when correct and creepy when wrong; transparency and controls are core design requirements.
Share
What it means
Persistent stores (profile fields, summaries, vector memories) re-injected into future prompts so the assistant “remembers” you.
Why designers should care
Users need to view, edit, delete, and scope memory (this project only vs global), especially after policy or team changes.
Example
A design copilot remembers your design system URL and preferred critique rubric; settings show each memory item with last updated and delete.
Common mistakes
- • Silent memory that users cannot inspect or reset.
- • Mixing personal and workspace memories without labels.
- • Remembering stale facts after major product rebrands or role changes.