A five-minute AI demo can look flawless, but the real test comes when customers push the product into workflows the builder never imagined. At TechCrunch Disrupt 2026, leaders from Anthropic, Gamma, and Clay will examine that gap on the AI Stage, asking why some AI deployments become essential while others remain experiments abandoned after a few tries.
Anthropic's Head of Applied AI, Cat de Jong, works directly with enterprises putting Claude into critical workflows. She sees the patterns across deployments: where they succeed, where they stall, and what separates organizations extracting real value from those still running pilots 18 months later. Gamma CEO Grant Lee and Clay CEO Kareem Amin bring the founder's view, pressure-testing those patterns against what they see in products people actually use.
Gamma has grown into a visual communication tool approaching 100 million users, giving Lee insight into what makes an AI product worth returning to. Clay, which was one of the launch apps integrated into Claude when Anthropic introduced interactive workplace tools, offers a close look at how AI fits into existing go-to-market workflows. The three speakers agree on the core challenge: moving from capability to dependable, everyday use is harder than building the demo. Their differing vantage points — one cross-enterprise, two product-level — should make the conversation useful for anyone selling AI into the enterprise.