Agentic Memory and Context in Data 360

Give your agents the ability to remember past conversations, learn user preferences, and retrieve relevant context automatically using Agentic Memory and Context in Data 360. By grounding agent responses in persistent memory and structured data, your agents provide continuity across sessions, personalize interactions, and make intelligent decisions based on complete customer understanding.

You can register agents, configure memory extraction and governance, and retrieve context using different modes depending on your data and interaction needs.

REQUIRED EDITIONS

Available in: All Editions supported by Data 360. See Data 360 edition availability.

Agentic Memory and Context provides your agents with these core capabilities.

  • Persistent conversation memory: Agents maintain awareness of what happened in past sessions, such as topics discussed, decisions made, and actions taken, without replaying full conversation transcripts.
  • Automated knowledge extraction: Data 360 extracts facts, preferences, and summaries from agent conversations at a periodic cadence to increase understanding of each individual over time.
  • Governed context delivery: The GetContext API assembles and delivers the right context to agents on demand and respects your org’s security model including object-level, field-level, and record-level access controls.
  • Multi-agent continuity: Because all memory is linked to a Unified Individual in your data space, users experience continuity regardless of which agent they interact with. Agents retain all context during handoffs.
  • Real-time availability: Conversations, session summaries, and profile memories become available for context retrieval within seconds of ingestion.

Consider these use cases for Agentic Memory and Context.

  • Multi-turn technical support: Agents remember prior conversations, steps already attempted, and error messages observed across sessions. This eliminates repeated explanations to accelerate resolution.
  • Sales conversations: Agents recall prospect preferences mentioned in earlier calls and surface relevant options immediately to create personalized and connected experiences.
  • Web agent interactions: Agents read browsing and search history to understand what shoppers are looking for and narrow results accordingly without requiring explicit input.
  • Agent handovers: When conversations transfer between agents, context flows seamlessly. The receiving agent is aware of previous discussions without requiring the user to repeat information.
  • Employee assistant scenarios: An internal HR agent remembers an employee’s prior leave requests, benefits selections, and ongoing accommodation needs without requiring them to re-explain their situation.

Agentic Memory and Context respects your org’s security model at every layer.

  • Object-level security (OLS): Controls which users can access which Data Model Objects.
  • Field-level security (FLS): Controls which fields within those objects are visible.
  • Record-level security (RLS): Controls which specific records users can view.
  • Agent-level access control: Determines which agents can read which memories.

All memory operations respect these governance rules. Context retrieval through the GetContext API automatically filters results based on the requesting user’s permissions. See Data Governance.