Webinar
From one agent to many: Building memory that scales
Build complex, multi-agent systems successfully.
A stand-alone agent can remember a conversation, call a few tools, and likely work once deployed. But building a system of agents is where things start to get complicated. With every additional agent, their relationships and connections grow and sprawl. That’s when the troubles begin.
In this session, we’ll walk through the problems that show up at scale, then build the solutions in a live demo.
Now, you can anticipate the common growth challenges that happen when building beyond agent one, and learn how Redis turns a sprawl of agents into a system that works together.
What you'll learn
- Scaling up: What changes the moment you grow beyond a single agent
- Shared memory: How to share memory across agents without mixing up context, using role-based access control (RBAC) to govern which agents can read and write what
- Conflicting facts: How memory should update when new information contradicts old information
- Seamless handoffs: Why agent-to-agent streams and handoffs need to flow smoothly from start to resolution
- Reliable persistence: How to build persistence so a restart never repeats an action
- Data deletion: How to delete sensitive data, and verify it’s actually gone
- Memory portability: What it takes to move memory when your stack changes
- Knowing limits: Where open source and single-agent memory solutions stop being (good) enough
Speakers

Redis
Karin Landers
Sr. Product Marketing Manager