Working AI: Glean Agents

In this episode of the "Working AI" podcast, Eddie Zhou, a founding engineer at Glean, discusses the concept of AI agents, differentiating them from LLMs and emphasizing the importance of enterprise-specific data and tool awareness for their success. He explains how Glean is scaling support for different query classes through robust evaluation methods and personalization, highlighting the need to parameterize systems to adapt to unique enterprise structures. Zhou also touches on future-proofing agents by building modular, LLM-agnostic systems focused on core problems like context selection, and underscores the critical role of search in providing agents with precise and relevant information for effective autonomous actions. The conversation highlights Glean's approach to building and improving AI agents for enterprise use, and promotes Glean's user conference, Glean Go, and an agents white paper for more information.