In today’s enterprise AI landscape, the lion’s share of effort (≈ 80 %) is not in tweaking the latest large language model, but in integration, orchestration and real-time data movement across systems. Meanwhile, only about 20% of the challenge sits in the model itself. This shift means that as we move from monolithic AI tools to truly autonomous, multi-agent systems, the architecture behind agentic AI matters more than ever.

In this talk, we will explore how the evolution from classic workflow automation to agentic AI systems demands a foundation built on event-driven architecture. Drawing on the lessons of the microservices era — where synchronous, point-to-point calls created brittle systems — we’ll see how EDA decouples agents, enables asynchronous coordination, and builds resilience at scale.

By the end of this workshop you’ll have:

A clear framework for when traditional workflow automation makes sense vs. when you should move to agentic, event-driven orchestration.

Practical understanding of how Solace Agent Mesh implements this architecture: event mesh, orchestrator, agents, gateways, data management, and observability.

Insight into how to focus 80% of your time on integration, event-driven design and data flows — not just the 20% of model fine-tuning — to make agentic AI production ready.

Actionable take-aways on how to design, implement and scale event-driven agentic workflows in your own organisation.