Automated tools can help insurers process claims faster, but lawsuits and regulators are testing whether speed has come at the expense of meaningful clinical review, clear disclosure and fair appeal rights.
A denial issued in seconds can leave patients spending weeks trying to prove that a qualified person reviewed their care.
A patient receives a health insurance denial after a routine claim is submitted. The notice says the treatment does not meet policy rules. It does not explain whether a clinician reviewed the patient’s record before the decision was issued.
That uncertainty is now central to the debate over artificial intelligence and automation in health insurance.
Insurers say automated tools can improve consistency, detect errors and process claims faster. Patients, physicians and plaintiffs’ lawyers argue that some tools can scale denials faster than affected households can understand or appeal them.
The concern is no longer theoretical. ProPublica reported that Cigna doctors rejected more than 300,000 claims over two months using an internal system known as PxDx, spending an average of about 1.2 seconds on each case. Cigna has disputed that characterisation, saying the system was used for certain administrative reviews and not to deny medical necessity. The dispute helped bring automated claim review into public and regulatory scrutiny.
Speed and scale are testing the meaning of clinical review
Health insurers have long used rules, codes and software to check claims. The new issue is whether automation is being used to support human review, or whether it is effectively substituting for it.
That distinction matters. A denial affects whether a patient receives care, pays out of pocket or enters an appeal process that can take weeks or months. If the patient cannot tell whether a qualified clinician assessed the medical facts, the right to appeal becomes harder to use.
Cigna’s PxDx system is one example. Reporting and litigation have alleged that claims could be rejected at high volume without doctors opening individual patient files. Cigna has said the reporting was incomplete and that the system was used for certain administrative reviews, not to deny medical necessity.
UnitedHealth has also faced litigation over its alleged use of the nH Predict algorithm in Medicare Advantage post-acute care decisions. In March 2026, a federal magistrate judge in Minnesota ordered UnitedHealth to produce broad discovery in a lawsuit accusing the insurer of using an AI algorithm to wrongfully deny post-acute care. UnitedHealth has denied the plaintiffs’ allegations and has argued that nH Predict does not make final coverage decisions.
Regulators are paying closer attention. A 2025 National Association of Insurance Commissioners survey found that 84% of responding health insurers used, planned to use or were exploring AI or machine learning. Among individual major medical insurers, 71% were using, planning or exploring AI or machine learning for utilisation management practices, and 68% for prior-authorisation approval processes.
California has gone further than many markets. Its Physicians Make Decisions Act, effective 1 January 2025, requires that a denial, delay or modification of health care services based on medical necessity be made by a licensed physician or qualified health care professional competent to evaluate the clinical issue. The law allows automated tools to assist, but not to replace, the required human decision-maker.
Patients can protect themselves with three practical checks
The first check is to ask for the basis of the denial in writing. Patients should request the policy provision, medical guideline, diagnosis code, procedure code and document list used to make the decision. A denial that cannot clearly identify the evidence used is harder to defend.
The second check is to ask whether a licensed clinician reviewed the case. The question should be specific: who reviewed the medical necessity issue, what qualification did they hold, and did they review the relevant clinical record? The answer may determine whether the appeal should challenge process as well as outcome.
The third check is to use the appeal window quickly. Health insurance denials usually have fixed deadlines. Patients should preserve the denial letter, medical records, physician notes, prior-authorisation history and all messages with the insurer. A treating physician’s written statement can be especially important where the denial depends on medical necessity.
Patients should also ask whether the decision was based on an automated tool, algorithm or third-party platform. Not every insurer will disclose this in detail, and disclosure rules vary by jurisdiction. But the question creates a written record and may help doctors, advocates or regulators assess whether the review process was fair.
Where the case involves urgent care, patients should ask about expedited appeal rights. A normal appeal timeline may not be appropriate when delay could affect treatment, discharge, rehabilitation or medication access.
Automation cannot replace accountability
AI and automated review tools may have legitimate uses in health insurance. They can flag missing data, detect duplicate billing, identify coding mismatches and help insurers manage large volumes of claims.
The consumer risk begins when speed becomes a substitute for judgment.
A denial should tell the patient what was rejected, why it was rejected, what evidence was considered and how to challenge the decision. If an insurer uses automation, the process still needs accountable human review where medical judgment is required.
A fast denial is not automatically unfair. A denial that no one can explain is.
"Don't hesitate anymore! Now, at BankQuality, Talk it all out with a relaxed mind about the services, behaviour and products of your banks! Review and Rate your Bank TODAY! #TalktoBQ — www.bankquality.com"