AI Explained: Verified Data is a Missing Piece of the Agentic Infrastructure Puzzle

Many enterprises are racing to put agentic AI into production. Very few have the data foundation to make it work. Agents don't just read data, they act on it in credit decisions, compliance checks, sales workflows, and supply chain calls. When the underlying data isn't verified or governed, autonomy becomes a liability. Join Gary Kotovets, Chief Data & Analytics Officer at Dun & Bradstreet, for a candid conversation about why verified data is where agentic AI reliability must start. What you'll learn: – Why does agentic AI deployments move faster when the data layer is already trusted? What does "trusted data" actually require to build? – What should governance frameworks look like when AI agents make autonomous decisions, not just surface recommendations? – How do data lineage and entity resolution keep multi-agent workflows accurate as agents hand off information across systems? AI Explained is our AMA series featuring experts on the most pressing issues facing Agentic and AI teams. 00:00 Introductions 05:30 Building the Foundation for Agentic AI 08:08 What Verified Data Really Means 10:43 Tackling Hallucination & Non-Determinism 15:08 Multi-Agent Workflows in Practice 19:08 Governance, Ethics & Human Oversight 21:51 Small Language Models & Cost Savings 27:26 Risk, Regulation & Data Provenance 37:30 Model Training & the Future 47:15 Rapid Fire and Closing