Why Insurance Companies Are Using AI to Deny Your Claims — And What You Can Do
Why Insurance Companies Are Using AI to Deny Your Claims — And What You Can Do
5 minute read
If your insurance claim was denied faster than any human could have read it, you are not imagining things. Insurance companies have quietly outsourced claim reviews to artificial intelligence systems that can process thousands of decisions per day, and the denial rates that come out of these systems are raising serious alarms for patients, doctors, and state regulators alike.
This is not a conspiracy theory. It is a documented shift in how coverage decisions get made — and understanding it is the first step toward doing something about it.
How AI Denial Systems Actually Work
Several large companies specialize in what the industry calls "utilization management." These are third-party vendors that insurers hire to review whether a prescribed drug or procedure meets coverage criteria. The biggest names in this space include eviCore (now part of Evernorth, which is Cigna's health services arm), Carelon (formerly AIM Specialty Health, now owned by Elevance Health), and Cohere Health.
These systems work by comparing your claim against a database of clinical criteria. When your doctor submits a prior authorization request, an algorithm scores the request against coverage rules. If the documentation does not precisely match the criteria the algorithm is looking for, the system flags it for denial or sends it to a human reviewer with a denial recommendation already queued.
A 2023 investigation by ProPublica and CBS News found that a UnitedHealth Group subsidiary used an AI model called nH Predict to issue blanket denials for post-acute care claims. The algorithm had a 90 percent denial rate for some claim types. Human reviewers, working from the algorithm's recommendations, spent an average of just 1.2 seconds reviewing each case before confirming the denial.¹
The problem is not that AI exists. It is that these systems are optimized for consistency and cost, not for the specific circumstances of your health situation.
Why Denial Rates Are Rising
Prior authorization denial rates have increased steadily over the past decade. A 2023 report from the American Medical Association found that 94 percent of physicians said PA requirements had delayed necessary care for their patients.² One in five patients abandoned their treatment entirely because of prior authorization barriers.²
The expansion of AI-driven review has accelerated this trend. When a human reviewer looks at a case, they can exercise judgment. When an algorithm does it, it looks for exact documentation matches. If your doctor's notes use slightly different clinical language than the algorithm expects, or if a single supporting document is missing, the system defaults to denial.
States are beginning to respond. California passed SB 1120 in 2024, which requires insurers to have a licensed physician review AI-generated denials before they become final.³ Other states are watching. But for now, if you are sitting with a denial letter, the regulatory landscape has not yet caught up to the technology being used against you.
What Documentation Actually Beats the AI
The good news is that AI denial systems have a weakness: they are rules-based. If you give them the right information in the right format, they often reverse. Here is what tends to make the difference in an appeal.
Clinical necessity language that matches the insurer's own criteria. Your doctor's appeal letter should use the exact clinical terminology from the insurer's coverage policy. Many of these policies are publicly posted on the insurer's website or can be requested directly. If the policy says "failed two prior agents from the same drug class," the letter needs to explicitly state that you tried those agents and what happened.
Documentation of what you have already tried. Step therapy requirements (more on this in another post) mean that many denials happen because the insurer believes you have not tried cheaper alternatives. A clear, dated history of prior treatments, doses, and outcomes is essential.
A letter of medical necessity from your prescribing physician. This letter should address the specific denial reason stated on your denial notice, not just make a general case for the drug. Generic letters of medical necessity are easy for AI systems to ignore. Letters that directly rebut the stated denial criteria are much harder to dismiss.
Peer-reviewed literature. If your drug was denied as "not medically necessary" or "experimental," citations to published clinical guidelines (such as those from the American Diabetes Association, the American College of Rheumatology, or other specialty societies) can shift the framing of your appeal significantly.
What This Means for You
If you received a denial, the first thing to do is request your denial notice in writing if you do not already have one. This document must, by law, state the specific reason for the denial and the criteria used. Read it carefully, because the path through an appeal runs directly through the stated denial reason.
Then gather your documentation: your treatment history, your physician's notes, any relevant lab results, and any prior authorization submission your doctor made. You are building a case, and every piece of evidence matters.
You have the right to appeal, and the numbers are more encouraging than you might think. Studies suggest that up to 80 percent of appeals that go through an external review process are decided in the patient's favor.⁴ The barrier is that less than 1 percent of people with denied claims actually file an appeal.⁵
Ellen can help you understand your denial reason and build an appeal letter that speaks directly to the AI's criteria. Start here →
Frequently Asked Questions
How do I know if AI was used to deny my claim?
You can ask your insurer directly whether an automated decision-making system was involved in your denial. Under some state laws and emerging federal guidance, insurers are required to disclose this. You can also request a copy of the clinical criteria used to evaluate your claim.
Can I appeal an AI-generated denial?
Yes. All insured individuals have the right to an internal appeal, and most have the right to an independent external review as well. External reviews are conducted by independent organizations unaffiliated with your insurer, which removes the AI system's influence entirely.
What is a peer-to-peer review?
A peer-to-peer review is a phone call between your doctor and a physician at the insurance company or utilization management vendor. Your doctor can make the case directly for your treatment. Requesting a peer-to-peer review is often one of the fastest ways to reverse a denial before a formal appeal is necessary.
Sources: