Skip to main content

What is User Memory?

User Memory lets your agent remember things about individual users across conversations — their name, preferences, past decisions, and more. Each user gets their own private memory.
User Memory vs Training Data: Training data teaches your agent what it knows (your FAQs, products, policies). User Memory personalizes how it responds to each individual user.

How to Enable

  1. Go to your agent’s Settings page
  2. Scroll to Features & Options
  3. Toggle User Memory on
Once enabled, your agent will automatically:
  • Learn important details from conversations (names, preferences, decisions)
  • Recall relevant memories at the start of each new conversation
  • Update memories when information changes

What Gets Remembered

Your agent automatically picks up on:
  • Facts — user’s name, company, role, location
  • Preferences — communication style, product interests, language
  • Decisions — choices made, plans discussed, agreements reached
  • Context — ongoing projects, past issues, relationship history

Managing Memories

View and manage what your agent remembers about each user:
  1. Go to the Subscribers page
  2. Click the memory icon next to a subscriber
  3. Browse, search, or delete individual memories
Each memory shows its content, type (semantic, episodic, or procedural), and when it was created.

Best Practices

✅ Enable for agents that have repeat users (support, coaching, sales) ✅ Review memories periodically via the Subscribers page ✅ Let users know your agent remembers context (builds trust) ❌ Don’t enable for anonymous one-off interactions (no user to associate memories with) ❌ Don’t rely on memory for critical data — it’s AI-extracted, not manually entered

Next Steps