The Agentic Gap: What Enterprises Think vs. What Actually Works With Jeff Dalton

About this episode

Jeff Dalton discusses the gap between enterprise expectations and the realities of AI implementation.

In this episode of AI Explained, we are joined by Jeff Dalton, Head of AI and Chief Scientist at Valence. Jeff has spent two decades at the intersection of research and industry, from building early conversational search benchmarks at Carnegie Mellon and Microsoft to leading the AI behind Nadia, Valence's purpose-built enterprise coaching assistant. He discusses the fundamentals of agentic system design that still hold from classical AI theory, why evaluation has to come before the prompt, how he approaches memory as a first-class object in coaching systems, and the defense-in-depth approach to guardrails that keeps complex agents safe across diverse enterprise deployments.

People in this episode

Guest: Jeff Dalton

Topics covered

• agentic system design
• AI theory
• enterprise coaching
• evaluation methods
• memory in AI
• safety in AI

Keywords

• agentic systems
• AI evaluation
• enterprise AI
• coaching systems
• memory in AI
• safety measures

Mentioned in this episode

Organizations: Valence, Carnegie Mellon, Microsoft

More episodes of AI Explained

• AI Agents Have an Identity Complex With Jeff Malnick · May 20, 2026 · 51 min
• Lessons from a Physician-CIO on AI Governance with Dr. Stacey Johnston · April 23, 2026 · 52 min
• How to Prevent AI Agents from Going Rogue With David Kenny · February 6, 2026 · 51 min
• Building Agents at Scale: Lessons from the Front Lines With Gary Stafford · September 25, 2025 · 58 min
• Lessons Learned from Building Agentic Systems With Jayeeta Putatunda · August 16, 2025 · 47 min
• Agent Wars: The Hype, Hope, and Hidden Risks with Nate B. Jones · July 24, 2025 · 56 min