Redpanda shares how they evolved Redleader, their internal AI customer experience system, from a single prompt-driven agent into a governed multi-agent architecture running on Redpanda's Agentic Data Plane (ADP). The revamped system uses specialized agents (Concierge, Records Custodian, Orchestrator, Sentiment Discriminator, Activity Monitor) that coordinate through durable event streams. These streams carry inbound messages, routing decisions, escalation signals, human overrides, and audit metadata, enabling structured collaboration, governance, and continuous learning across the agent ecosystem.
Table of contents
Creating a multi-agent habitat #Agents that “think” before they act #Ready to build? #Sort: