THE ASK
Can AI help payers and other healthcare stakeholders flag the first signs of rising medical costs?
Rising medical costs are a significant headwind for healthcare payers and other stakeholders across the sector. A substantial portion of the forward-looking information relevant to cost drivers emerges in qualitative form, including regulatory draft proposals, clinical trial outcomes, policy commentary, reimbursement negotiations, and industry guidance. Historically, domain experts have interpreted these dispersed information sources and translated them into implicit judgments about future cost trends. But this expert-led process is difficult to scale, systematize, or replicate within a quantitative framework. A research team serving the healthcare industry needed a more efficient way to capture these signals sooner and turn them into consistent, actionable insights on medical cost trends.
THE APPROACH
Milliman developed an AI-powered system to extract and analyze healthcare cost signals from news at scale
Leveraging large language models, Milliman built a three-part custom AI solution:
- The system continuously extracts news from more than one million sources across mainstream media, healthcare industry publications, and data vendors.
- Using these news articles, Milliman’s proprietary sentiment index draws on the firm’s deep healthcare expertise to assess how different industry developments might affect prices for prescription drugs, outpatient care, and other inputs.
- An interactive dashboard lets users visualize these sentiment trends, request AI-generated narrative explanations, and query the full dataset of news results through a conversational interface.
THE OUTCOME
Real-time signs of rising medical costs, and empowered decision making for the healthcare sector
Milliman’s sentiment tracker gave the research team access to a new range of industry insights. The first signs of rising medical costs were now apparent in real time—captured from news as it broke, not weeks later when it appeared in claims data. When the index moved, analysts could immediately request an AI-generated narrative explaining what was driving the shift and which companies were most exposed. And by querying the full historical dataset through a chatbot interface, teams could investigate past events and understand how similar dynamics had played out before—turning a vast archive of news into a practical risk intelligence tool for empowered healthcare financial decisions.
