{"author_name":"Voices of Search // A Search Engine Optimization (SEO) & Content Marketing Podcast","author_url":"https://art19.com/shows/voices-of-search/episodes/e90c8549-781e-460a-bb20-d7b21fd6be46","description":"<p>Agentic AI workflows fail without classified first-party data. Venkata Pagadala, AI Product Manager for Search at AT&amp;T, classified 5.8 million queries, removed 3 million noise queries, and uses the resulting persona-based insights to build mockups that win leadership buy-in. He outlines a competitor-navigation and Wikipedia-driven L1/L2/L3 taxonomy that surfaces zero-search-volume, high-intent demand from Reddit, Quora, and YouTube, plus a UGC-based research framework that treats reviews, ratings, and brand response rates as primary inputs. He also covers persona-based query fan-outs as a replacement for single-prompt tracking, and documentation systems that let AI agents repeat classification at scale.</p>","html":"<iframe src=\"https://art19.com/shows/voices-of-search/episodes/e90c8549-781e-460a-bb20-d7b21fd6be46/embed\" style=\"width: 720px; height: 200px; border: 0 none;\" width=\"720\" height=\"200\" scrolling=\"no\" sandbox=\"allow-scripts allow-popups allow-popups-to-escape-sandbox\"></iframe>","provider_name":"ART19","provider_url":"https://art19.com","title":"Building AI Agents That Actually Work for Enterprise SEO","type":"rich","version":"1.0","width":720,"height":200}