Agentic AI Fails: Why Insurers Miss Returns

Insurance firms using agentic AI may not see financial returns if they treat it like a simple software rollout. This is according to Dr. Magdalena Ramada Sarasola of WTW. Here's the thing: many executives mistake agentic AI for just a powerful chatbot. But the real issue is where companies believe the value truly comes from. What's interesting is that the design choices for an AI model are more critical than the model itself. Orchestration and contextualization consistently outperform raw model intelligence. The way you structure an agent's access to tools, the order it uses them, and the guardrails you implement are key. Embedding institutional expertise into the system is where competitive advantage is found. Two insurers using the same foundational model can get very different results based on their architecture. The bottom line: While general-purpose AI can boost individual productivity, financial impact comes from specific, insurance-focused AI systems deeply integrated into core workflows. This matters because understanding this distinction can help insurers achieve actual returns from their AI investments.