Sponsored by: Pogo | You're Overthinking It: What Analyzing $500B in Consumer Spend Taught Us About Building AI Agents
Overview
| Experience | In Person |
|---|---|
| Track | Artificial Intelligence & Agents |
| Industry | Enterprise Technology, Retail & Consumer Goods, Financial Services |
| Technologies | AI/BI, Databricks Apps, Agent Bricks |
| Skill Level | Beginner |
At Pogo, we built AI research agents to help brands analyze hundreds of billions in consumer spending data. Our biggest lesson: the breakthrough came after we threw most of it away. We started with a complex system of specialized agents. It was fragile, unreliable, and impossible to iterate on. When we simplified to a single agent, everything changed. In this session, we’ll share why simpler agent architectures dramatically outperform complex ones, what the frontier AI labs’ own designs reveal about where the industry is heading, and how this approach powers Pogo’s AI research agent - where brands analyze purchase data from 3M Americans and interview thousands of verified buyers through an AI video interviewer, with results in hours.
Session Speakers
Luke Morgan-Scott
/Head of Data
Pogo