Skip to main content

Memories

Overview​

Memories are workspace-scoped knowledge snippets that help Attive stay consistent across conversations. They’re designed for information you want the assistant to reliably remember and reuse—like definitions, customer context, go-to-market strategy, and “how we do things here.”

Usage​

General usage​

  • Use memories to ground answers in your workspace context.
  • Let Chat auto-recall relevant memories as you ask questions (in default mode), or pin specific memories when you want to be explicit.
  • Use tags and categories so people can find and reuse the right memories.

Creating and managing memories​

Memories are managed in the workspace Context / Library area.

What you can do:

  • Create memories for facts or guidance you want to persist.
  • Edit and refine memories as your business changes.
  • Organize memories with tags and categories.
  • Control what’s active by enabling/accepting memories and rejecting ones you don’t want used.

Memory types (high level):

  • Manual: intentionally created by a user.
  • Auto: generated from activity (for example, extracted from conversations) and used when relevant.
  • Always: baseline context that should always be considered.

Using memories in Chat​

Memories show up in Chat in a few different ways depending on how you’re working.

  • Auto-updating conversation context (default mode)

    • As you ask questions, Chat can automatically recall relevant memories and add them to the conversation’s active context.
    • If you want the context to stay fixed, you can lock it.

  • Explicit memories (pin what matters)

    • You can explicitly select memories for the next message (from the composer’s context picker).
    • You can also add memories directly into the conversation context from the Memory sidebar.

  • Review what was used

    • For a given message, you can inspect which memories were recalled.
    • You may also see suggested memories you can accept (to keep) or reject.

  • Agents change memory behavior

    • If you select an agent, the agent’s pre-configured memories become the base context.
    • In agent mode, Chat does not keep expanding memory context automatically.

Examples​

  • Company context

    • “We sell to mid-market SaaS. Our ICP is RevOps leaders at 200–2,000 employee companies.”

  • Definitions and rules

    • “When we say ‘healthy pipeline’, we mean opportunities with next steps scheduled within 14 days.”

  • Messaging guidance

    • “When writing outbound, keep it under 120 words and lead with a single quantified insight.”

Features​

FeatureWhat it does
Workspace knowledgeStores reusable context that multiple people and chats can rely on.
Retrieval (relevant recall)Brings the most relevant memories into a conversation when needed.
Conversation memory contextMaintains an active set of memories for a chat, which can auto-update or be locked.
Explicit selectionLets you choose exactly which memories should apply to the next message.
GovernanceSupports accepting/enabling, rejecting, and refining memories over time.
OrganizationUses tags and categories to keep memories discoverable and reusable.