Building Enterprise-Scale Agentic Claims Automation and AI Observability With Databricks

As enterprises move from GenAI experimentation to production, the hardest problems are no longer models or agent frameworks—they are enterprise‑scale orchestration and observability.At Suncorp, we have deployed 16 production‑grade agentic AI solutions that are now converging into an enterprise claims automation capability. This session shares the reference architectures and engineering patterns behind this journey, focusing on two pillars: agentic orchestration, where Databricks supports low‑latency, governed workflows and shared agent memory using Delta Lake, Vector Search, Agent Bricks, and Lakebase, and AI observability at scale, where we use MLflow evaluation, OpenTelemetry, Databricks Apps, and AI/BI Dashboards to trace agent behaviour, decisions, quality and risk.Attendees will leave with practical patterns for building orchestrated, observable agentic AI in a regulated enterprise.