# Interview: Half the Tokens: Giving AI Agents Memory | William Lyon, Neo4j

**URL:** <https://community.neo4j.com/t/interview-half-the-tokens-giving-ai-agents-memory-william-lyon-neo4j/81008>\
**Category:** General\
**Created:** [September 10, 2026, 12:41pm UTC](https://community.neo4j.com/t/interview-half-the-tokens-giving-ai-agents-memory-william-lyon-neo4j/81008 "2026-09-10T12:41:33Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Ari\_Neo4j](https://sea1.discourse-cdn.com/flex021/user_avatar/community.neo4j.com/ari_neo4j/32/34352_2.png) [@Ari\_Neo4j](https://community.neo4j.com/u/Ari_Neo4j)\
**Post date:** [September 10, 2026, 12:41pm UTC](https://community.neo4j.com/t/interview-half-the-tokens-giving-ai-agents-memory-william-lyon-neo4j/81008/1 "2026-09-10T12:41:33Z")

</div>

![composed (1)](https://us1.discourse-cdn.com/flex021/uploads/neo4jcommunity/original/3X/3/c/3c72511ef6dd7278402a0aa4c15c8e255ec0dfdb.jpeg)

Some of you know @lyonwj and many of your don't... Will has been thinking about data, systems, and graphs since long before the current AI wave.

Recently, Will spoke with **[Sri Kalimani](https://www.linkedin.com/in/sri-kalimani/), Technology Product Manager & Podcast Host, on [The Foundry Alley](https://www.youtube.com/@foundryalley/videos)**

👉 Watch the conversation and join the discussion in the Neo4j Developer Community: [https://dev.neo4j.com/will\_lyon\_foundry\_alley](https://dev.neo4j.com/will_lyon_foundry_alley)

Learn how Will got involved with Neo4j and Graph Technology and all of the critical topics listed below.

**Chapters**

[00:00](https://www.youtube.com/watch?v=7a6godUqPSQ) Introduction to Neo4j and Graph Technology  
[01:07](https://www.youtube.com/watch?v=7a6godUqPSQ&t=67s) The Role of Knowledge Graphs in AI  
[02:16](https://www.youtube.com/watch?v=7a6godUqPSQ&t=136s) Applying Knowledge Graphs to Generative AI  
[03:23](https://www.youtube.com/watch?v=7a6godUqPSQ&t=203s) Cost Control and Efficiency in AI Agents  
[05:20](https://www.youtube.com/watch?v=7a6godUqPSQ&t=320s) Data Storage and Retrieval Changes Post-AI  
[08:47](https://www.youtube.com/watch?v=7a6godUqPSQ&t=527s) Handling Unstructured Data with Vectors and Graphs  
[11:38](https://www.youtube.com/watch?v=7a6godUqPSQ&t=698s) Scaling GraphRAG for Large Data Sets  
[25:11](https://www.youtube.com/watch?v=7a6godUqPSQ&t=1511s) Demo: Neo4j Agent Memory Service in Action  
[36:42](https://www.youtube.com/watch?v=7a6godUqPSQ&t=2202s) Best Practices for Building Agent Memory Systems  
[45:31](https://www.youtube.com/watch?v=7a6godUqPSQ&t=2731s) The Impact of Graph Memory on Enterprise AI

What should AI agents remember between runs?

Explore Neo4j Agent Memory: [Neo4j Agent Memory - Neo4j Agent Memory](https://neo4j.com/labs/agent-memory/)