Every time a physician submits a prior authorization request, someone has to interpret the rules. Those rules live in policy documents, provider contracts, and clinical guidelines that health plans update regularly but rarely make computable. Different staff members read the same clause differently. The same request gets handled two ways by two departments on the same day. Patients wait. Decisions vary. The inconsistency is not the result of negligence. It is the result of a system whose underlying logic was never made readable enough to be applied uniformly.
Boost Health AI, a Chicago company that launched in October 2025, is attempting to change that. Its argument is straightforward: before health plans can use AI responsibly, they need to be able to read their own rules.
The Scope of the Problem
The administrative burden on American healthcare is well-documented. According to a 2024 survey by the American Medical Association, physicians spend an average of 13 hours per week processing prior authorization requests, with 40% of practices employing staff dedicated exclusively to that task. Twenty-nine percent of physicians report that authorization delays have produced serious adverse events for patients in their care.
Health plans are under significant financial pressure simultaneously. The National Association of Insurance Commissioners reported that health plans’ aggregated net income fell 14% in the first half of 2024, while cash flow from operations dropped 86% over the same period. McKinsey estimates that payers adopting AI and automation could reduce administrative expenses by 13 to 25%. The financial incentive to automate is clear. The risk is automating the wrong things, or automating without the transparency needed to catch errors before they reach patients.
A 2024 U.S. Senate investigation found that Medicare Advantage insurers used algorithmic tools to deny patients post-acute care at rates that drew significant scrutiny from lawmakers and patient advocates. The lesson from that episode was not that AI should be kept out of coverage decisions. It was that AI operating without traceable, auditable logic creates liability that eventually lands on real people.
What Boost Is Actually Building
Boost’s core product, the AI Foundry, converts policy documents, provider contracts, and clinical guidelines into machine-readable, auditable intelligence. AI Agents extract the decision logic from those documents. That logic then feeds into pre-built workflow templates covering claims processing, utilization management, care management, and compliance validation. Every decision the system produces carries citations, versioning, and a traceable audit trail.
The company’s commercial model is also worth noting. Rather than selling access to a hosted platform on a subscription basis, Boost licenses its technology as perpetual IP that the health plan owns and controls within its own environment. Clients can deploy on their own cloud, on-premise infrastructure, or through managed hosting.
“We believe payers should have sovereignty over their intelligence,” said Raheel Retiwalla, the company’s Chief Product Officer. “When you own the framework, you can evolve it as technology changes. You’re not waiting on someone else’s roadmap.”
An Early-Stage Company in a Fast-Moving Market
Boost is seven months old and has not disclosed named clients, funding figures, or revenue. Its first deployments have targeted care management and utilization management workflows. Early results, according to the company, include faster turnaround times and better coordination between clinical and operational teams. Those claims have not been independently verified.
The market it has entered is large and growing. The global AI for healthcare payer market is projected to expand from $5.7 billion in 2025 to $46.67 billion by 2035. Competitors include well-resourced incumbents like Oracle Health and Microsoft Nuance alongside a growing number of funded startups. Boost’s ownership model is a genuine differentiator, though whether large health plans will embrace the operational responsibility that comes with owning AI infrastructure rather than renting it remains an open question.
“Payers don’t need another application,” Retiwalla said. “They need an operating system for intelligence that helps their existing systems work together cohesively.”
The payer AI market is moving fast, under real regulatory pressure and real financial strain. Boost’s premise, that you cannot govern what you cannot read and cannot own what someone else controls, is grounded and worth taking seriously. Whether a seven-month-old company can deliver on it at enterprise scale is what the next few years will reveal.