R1 Buys Humata Health to Automate the Prior Authorization Headache With AI

R1, a large revenue-cycle management company, is acquiring Humata Health, a startup using AI to handle insurance prior authorizations — the approvals providers must secure before many treatments, and among the most-cited pain points in US healthcare.

Financial terms were not disclosed; the deal is expected to close in the third quarter of 2026.

What prior authorization is

Before certain treatments — expensive drugs, imaging, procedures, hospital admissions — a provider must obtain the insurer’s agreement to pay. Without it, the claim is denied after the fact and someone absorbs the cost.

The process means identifying the insurer’s policy for that treatment, assembling documentation showing the patient meets criteria, submitting it, and following up — frequently through appeals, and frequently through fax and phone rather than any modern system.

The insurer’s justification is cost control: preventing unnecessary or inappropriate care before it is delivered. The provider’s complaint is that the process consumes staff time on a scale out of proportion to what it catches, and delays treatment for patients who plainly qualify.

Why it is genuinely hard to automate

The difficulty is that there is no single system. Each insurer maintains its own policies, its own criteria and its own submission requirements, and those change continually.

A practice treating patients across a dozen plans faces a dozen sets of rules for the same procedure. Staff learn them through experience, and that knowledge is difficult to codify because it is scattered across policy documents, portals, and institutional memory about what each insurer actually accepts.

That combination — unstructured rules, high volume, repeated patterns — is a reasonable fit for language models, which is why the category has attracted so much attention.

What Humata does

Founded in 2023 by a Mayo Clinic-trained physician, Humata identifies insurer policies, builds authorization requests and shepherds them through approval — reporting a 96% first-pass approval rate.

That figure is the selling point and deserves scrutiny. First-pass approval means the request was accepted without appeal or resubmission, and it depends heavily on which requests are attempted — a system submitting only straightforward cases would post a high rate while leaving the difficult ones to staff.

Taken at face value it indicates the system is assembling documentation that satisfies insurer criteria the first time, which is where most of the wasted effort in the current process sits.

Why R1 wants it

R1 processes more than 600 million transactions a year across about 1,000 providers, and prior authorization is a natural extension of its claims-automation business.

“Humata significantly enhances our coverage of the authorization process,” said R1 CEO Joe Flanagan.

The logic is straightforward. Revenue-cycle management already handles the back end — submitting claims, chasing denials, collecting payment. Prior authorization sits at the front, and denials frequently trace to authorization problems, so a company managing the consequences has an obvious interest in the cause.

Scale also matters for this technology specifically. A system learning insurer behaviour improves with volume, and 600 million transactions is a substantial training and validation base.

The regulatory backdrop

Rules are tightening. A 2024 CMS rule mandates faster, electronic decisions, and insurers made voluntary commitments to deliver real-time responses for 80% of electronic prior-authorization approvals in 2026.

Both push in the same direction: authorization moving from a slow manual exchange to an electronic transaction with expected turnaround times. That favours providers with systems able to submit structured requests automatically, and disadvantages those still working by fax.

It also explains the timing of the acquisition. A regulatory shift creating demand for automation is a good moment to own an automation vendor.

The uncomfortable dynamic

Worth naming: this is an arms race rather than a solution.

Providers are deploying AI to submit authorization requests more effectively. Insurers are deploying AI to review them — and have faced litigation and scrutiny over automated denial systems. The likely outcome is automated systems negotiating with automated systems over whether a patient receives treatment.

That may reduce administrative cost on both sides, which is a real gain. It does not address the underlying question of whether this volume of pre-approval is warranted at all, and it entrenches the process by making it cheaper to operate.

What it means for patients

Patients rank prior authorization among the biggest obstacles to care, and providers cite added paperwork and treatment delays.

What the consolidation pattern suggests

A large incumbent buying a three-year-old AI startup is a pattern now repeating across health technology, and the logic on both sides is worth naming.

The startup has capability and no distribution. Selling into health systems is notoriously slow — long procurement cycles, integration with entrenched electronic record systems, and buyers who have been disappointed by previous vendors. A company with strong technology can spend years reaching a fraction of the market.

The incumbent has distribution and comparatively little capability in newer techniques. Building internally means competing for scarce talent against better-paying employers and moving at the speed of a large organisation.

Acquisition resolves both, and it is why relatively few of these companies reach scale independently. The consequence for buyers is that capability arrives bundled into existing vendor relationships rather than as competitive alternatives — which is efficient, and reduces the pressure on incumbents that independent competitors would apply.

Faster authorization is a genuine benefit if it means treatment starts sooner. The gain is administrative rather than clinical, and the patients most harmed by the current system are those whose requests are denied rather than delayed — a group automation on the provider side helps only insofar as better documentation changes the answer. Business news, not investment advice.