Most organizations have invested years in their data platforms — yet find them surprisingly unprepared for AI. Clean dashboards and reliable pipelines don’t automatically translate into systems that LLMs and agents can reason over reliably. This talk unpacks what it actually takes to make a data platform AI-ready.

Drawing on petabyte-scale work at Zendesk, the session covers the foundational shifts required: building semantic layers that ground models in trustworthy, governed metrics; designing for retrieval through entity extraction, vectorization, and semantic search; and introducing agentic automation that turns slow, manual data workflows into fast, self-service ones. We’ll explore why semantic context — not just raw data — is the missing link for AI, and how to architect for it without rebuilding from scratch.

Attendees will leave with a practical framework for assessing their own platform’s AI-readiness and a roadmap for closing the gaps that matter most.