Health insurers have invested heavily in chatbots, robotic process automation, and self-service portals, but a sponsored analysis from Infosys argues that these tools are deployed in silos. Friction remains at the handoffs between call centers, prior authorization units, claims, and compliance. The source cites J.D. Power data showing only 30 percent of commercial members see their insurer as a trusted partner, and an AMA survey finding that prior authorization costs physicians and staff an average of 13 hours per week, with 95 percent saying it delays care.
Agentic AI, unlike scripted automation, can pull data from multiple sources, identify missing documentation, and initiate follow-up across departments. In member services, it assembles eligibility, claims history, and prior interactions into one view before a call connects. In prior authorization, it checks documentation against payer criteria and only escalates ambiguous cases to clinicians. The source also points to financial stakes: hospitals spent roughly $43 billion in 2025 pursuing insurer payments, and CMS reported billions in improper payments for Medicare and Medicaid, often tied to documentation gaps.
For responsible deployment, the article emphasizes human oversight, clear escalation paths, bias monitoring, and audit trails. It highlights Highmark Health's collaboration with Abridge, where a system compares authorization requirements with real-time visit information and prompts physicians to collect missing details. The broader lesson, the source argues, is to start narrow and keep clinical and coverage judgment with humans.