Democratizing Marketing Analytics With Governed AI Agents on Databricks | Data + AI Summit 2026

What if marketers could answer complex questions without filing a ticket, waiting on data engineers, or knowing a single line of SQL? In this session from Databricks Data + AI Summit, Ankur Jain (Chief Cloud & Data Modernization Officer, Acxiom) and Bhaskar Dutta (Director of AI Product Management, Prophecy) show how agentic AI is collapsing a multi-week marketing analytics workflow into minutes. You'll see a live demo of how Acxiom and Prophecy partnered to build a self-serve marketing analytics platform on Databricks. The platform lets marketers generate audience segments, inspect governed data pipelines, and optimize campaign strategy through natural language — no data engineering handoffs required. Key topics covered: • Why traditional marketing analytics pipelines are too slow and skill-heavy for today's business pace • How Prophecy's context graph enables accurate, governed AI output • The "generate, refine, deploy" workflow for building marketing audiences with AI agents • How identity resolution and data enrichment form the foundation for trustworthy AI • Live demo: identifying high-propensity trade-in customers, building scored audiences, and optimizing campaigns for revenue lift • The full architecture: Databricks as the data fabric + Axiom for identity enrichment + Prophecy for conversational AI 🔗 Learn more about Prophecy: https://www.prophecy.ai 🔗 Learn more about Axiom: https://www.acxiom.com/ Timestamps 00:00 Introduction and disclaimer 00:54 Session overview: democratizing marketing analytics 02:06 Marketer’s problem: identifying at-risk customers 03:12 Why this requires data, engineering, and marketing alignment 04:12 Traditional data architecture and identity resolution 05:27 Why the current process is slow and complex 07:11 The case for self-serve, AI-powered analytics 08:41 Prophecy demo intro 10:08 Audience building 12:08 Data foundation example 14:04 Explore live data 15:45 Why governance matters 17:16 Context graph explained 18:38 Propensity model example 19:54 Generated pipeline 20:40 Deploying the app 22:06 Campaign optimization 23:20 Closing summary