Beyond Recall: Creating Agentic Personas with Persistent, Evolving Memory
summary
Claims that Neo4j becomes a long‑term memory layer for agents, turning short‑lived chat into a Neo4j graph-native memory that stores nodes like user, conversation, message, preference, ticket, event, decision and tool call. The system orchestrates LLM prompts with Neo4j reads and writes, supports multi-hop recall (for example customers whose renewals are nearby and had escalations and mixed sentiment last month), performs summarization and agent self‑reflection, and enforces patterns like separate logs from curated memory, time/source/confidence tracking, and controlled forgetting so personas evolve over weeks and months.
AI-written summary · about
chapters (13)
- 0:00Introduction and speaker intro
- 1:06Problem with agent memory loss
- 2:54Human vs AI memory storage
- 6:47Five factors of agent memory
- 8:12Limitations of traditional memory
- 9:14Why knowledge graphs matter
- 10:59Graph structure inspiration
- 12:46Neo4j memory system workflow
- 13:49Failure stories in agent memory
- 15:34Financial data use case
- 16:19Enterprise AI focus shift
- 17:21Benefits of persistent memory
- 19:10Closing thoughts and questions
related talks
28mAgentic Memory
15mBuilding Evolving AI Agents Via Dynamic Memory Representations Using Temporal Knowledge Graphs
51mMemory Masterclass: Make Your AI Agents Remember What They Do!
18mArchitecting Agent Memory: Principles, Patterns, and Best Practices
54mThe Anatomy of Memory in Humans and AI Agents
31mIs My Brain a Knowledge Graph? A Human Take on Connecting the Dots
23mConnecting the Dots With Context Graphs
15mMemory-Augmented AI Agents: Enhancing Long-Term Reasoning and Context Awareness