Learn how predictive modeling and machine learning appear in the insurance industry. This includes the journey of the first business question to a model that actually runs in production. I’ll start with business pain points insurers care about (risk scoring, claims triage, fraud detection etc.) and initially zoom in on the translation of the problem. As a business need moves through  the corporate hierarchy, the problem gets reframed, narrowed, and measured differently, until it becomes a technical prediction problem with a specific target, time window, and definition of “success.” The talk follows the full lifecycle of a technical problem from turning messy data into usable signals, iterating advanced algorithms under constraints and validating results in a way stakeholders can trust.