Most agentic AI demos work because the demo environment is perfect. Data is static. There is no back-pressure, no retries, and no real ownership boundaries.

Production is different.

In the real world, agents rarely fail because LLMs can't reason. They fail because workflows become brittle. Context gets lost. State becomes stale. Services change. New capabilities must be inserted. Load spikes unpredictably.

This workshop teaches participants how to build event-driven, production-level agentic systems using the Solace Agent Mesh. In this model, agents behave like microservices: autonomous, loosely coupled, independently scalable, and resilient to failure.

Participants will learn how the Solace Agent Mesh coordinates agent ecosystems through real-time events, supports dynamic agent discovery, and integrates agents using A2A and MCP protocols. The result is an architecture that evolves without rewiring pipelines and survives real enterprise conditions.

By the end of the workshop, participants will not just have an agent. They will have a working Agent Mesh: a distributed, orchestrated fabric that allows agents to collaborate across live enterprise environments in a safe, observable, and scalable way.