Finally Built the Training I Wish Every Front Desk Had See the Course →

The Robots Are Here, and They Are Denying Everything…

Tags:

Monica, the lead biller, hits submit on a clean claim.

CPT 99214.

Office visit.

Jack Smith has a stomach ulcer, confirmed by the physician.

Chart note is complete.

Everything is signed.

Coding is tight.

Submitted at 8:01 AM.

At 8:01 and 5 seconds, the denial lands.

“Documentation does not support medical necessity.”

Monica stares at the screen like it’s going to explain itself.

It doesn’t.

Meet Cortex. Your New Denial Machine.

Chad from IT told everyone last week that Cortex by Change Healthcare would “streamline denials” and “reduce turnaround time.”

Translation: insurance companies installed a real-time AI denial engine, and now Monica is losing to a bot that reviews claims faster than she can click “print.”

Of course, Cortex is owned by UnitedHealthGroup.

Because what isn’t owned by United at this point?

They manage insurance, authorizations, pharmacy benefits, billing platforms, and now AI-based denial systems.

If they acquire Amazon next, Jack might be getting his ulcer meds via Prime UnitedHealth.

Holly Still Thinks AI Means Fewer Denials.

Holly, the administrator, walks into Monica’s office holding a printed PDF from a vendor rep.

“It states here Cortex is trained on millions of claims to improve accuracy,” she says.

Monica doesn’t look up.

“Accuracy for who?”

Because in the last 48 hours, Cortex denied 19 claims with the same diagnosis and note structure that had been paid every week since 2021.

What changed?

Nothing on their end.

But clearly, Cortex decided something was wrong, and the system obeyed.

Jack Just Wants to Know Why His Bill Is So High.

Jack walks up to the front desk holding a letter from his insurance company.

“This says my visit wasn’t medically necessary,” he tells Jill.

“Am I going to get a bill?”

Jill freezes.

She has seen this look before.

She calls Monica over from the billing department.

Monica glances at the letter and immediately recognizes the phrasing.

“This is a denial from your insurance. You are not responsible for the balance. We have to appeal it.”

Jack raises an eyebrow.

“But the doctor said I had an ulcer.”

“You do,” Monica says.

“But the AI system they use flagged the note as not meeting their internal criteria. It’s not about the care. It’s about how their algorithm reads the documentation.”

She watches Jack process that.

The care was real.

The diagnosis was real.

But the payment system? Decided otherwise.

The Problem Isn’t the Documentation. It’s the Machine.

Cortex runs claims through an algorithm trained on historical data and payer policies.

But if your chart note does not match what it expects to see, the claim is auto-denied.

No human reads it first.

No appeal flag is triggered.

You are told to resubmit or appeal after the fact, while the system continues to run as if nothing happened.

Monica sends the exact same claim again the next day.

It gets paid.

Same code.

Same note.

Same everything.

And That Is Exactly the Point.

UnitedHealthcare did not build Cortex to help your billing team.

They built it to automate rejection.

They built it so no one has to look a patient in the eye and say, “We denied this for no good reason.”

Because now the machine does it for them.

No accountability.

No logic.

Just denials at scale.

The Bottom Line?

In 2026, insurance companies are no longer reviewing claims.

They are filtering them through AI and letting the machine decide.

Your chart note might be perfect.

Your code selection might be clean.

Your physician might follow every guideline.

And the claims still get denied.

Because Cortex said so.

No human reviewed it.

No real explanation followed.

No one warned you that the rules changed.

By the time you figure it out, Jack already has the denial letter.


Discover more from Sasha Exposes Healthcare.

Subscribe to get the latest posts sent to your email.

Leave a Reply

Discover more from Sasha Exposes Healthcare.

Subscribe now to keep reading and get access to the full archive.

Continue reading