Anthony R. O’Donnell // January 14, 2026 
(Image credit:Andrew Svk/Unsplash..)
For much of the past two years, the insurance industry’s conversation about generative AI has focused on experimentation at the edges: chatbots, document summarization, and isolated productivity tools. What has remained harder—and far more consequential—has been enabling AI systems not merely to reason about insurance data, but to act on it inside the core systems where underwriting, policy administration, and servicing actually occur.
That challenge is now coming into sharper focus with the emergence of the Model Context Protocol (MCP), a new open standard designed to let AI agents