AI Agents: What’s Working, What’s Not, and Why It Matters

Autonomous AI agents are dominating the narrative around the next phase of automation, yet real-world deployments tell a more mixed story. In practice, teams are encountering fragile reasoning chains, shallow contextual grounding, and rapidly escalating compute costs as systems scale. The panel convenes practitioners who are actively running agentic systems in production to distinguish hype from reality—exploring where agents deliver durable value, where they break down, and the design choices that determine the difference.