Why Healthcare Providers Are Spending Billions Just to Get Paid

Getting paid in healthcare should be simple.

It isn’t.

Providers are now spending over $25 billion a year just working through claims. Not because the claims are wrong, but because the process is.

About 15% of claims get denied upfront. That sounds reasonable until you realize most of them are eventually approved and paid anyway.

That means the system is not catching bad claims. It is slowing down good ones.

Every denial triggers the same loop. Review. Resubmit. Wait. Repeat. Weeks turn into months. Multiply that across thousands of claims and the cost explodes.

This is not a technology issue. It is a workflow problem built into the system itself.

Then there is prior authorization. Providers ask for approval before care. They get it. Then the claim can still be denied later.

So now they ask for permission, receive permission, and still have to fight to get paid.

That is not control. That is duplication.

The impact is immediate. Hospitals are operating with tighter margins than they have in years. Cash is delayed. Staff is tied up in administrative work. Investment in care gets pushed out.

At the same time, these inefficiencies drive up costs across the system, contributing to higher premiums.

Nothing about this is accidental. The system is doing exactly what it was built to do.

The problem is what it was built to do no longer makes sense.

Most denied claims end up getting paid. Billions are spent proving what was already true.

Until payer rules are standardized, unnecessary approvals are reduced, and claims are validated before submission, this does not change.

Providers will keep funding a process that works against them.

AI in Payer Operations: Efficiency Tool or Legal Liability?

The payer industry is moving fast on AI.

Claims are being automated. Prior authorizations are being streamlined. Risk adjustment is being augmented. Call centers are being replaced with conversational models.

The story everyone is telling is simple.
AI drives efficiency. Efficiency drives margin.

That story is incomplete.

What’s actually happening is this:
AI is moving faster than the controls required to manage it.

And that gap is where the real risk sits.

AI is no longer a support tool. It’s embedded directly into decision-making.

It determines whether a claim is paid.
It influences whether a prior authorization is approved.
It flags what gets reviewed and what gets ignored.

That shift matters.

Because once AI starts making decisions, you’re no longer optimizing workflows.
You’re automating judgment.

And most organizations are not set up to govern that.

There’s a problem building under the surface that few teams are willing to say out loud.

First, accountability starts to break down.

When a decision is driven by an algorithm, ownership becomes unclear.
Was it the plan? The vendor? The model?

In a manual process, responsibility is obvious.
In an automated one, it fragments.

Second, explainability becomes a real issue.

It’s easy to say a model flagged something.
It’s much harder to explain why in a way that stands up to audit, appeal, or legal review.

If you can’t clearly defend a decision, the efficiency you gained becomes irrelevant.

Third, and most important, mistakes scale.

A human makes errors one at a time.
AI makes them thousands at a time.

If the logic is flawed, the impact isn’t contained. It compounds quickly and quietly.

By the time it’s discovered, the exposure is already material.

This is where the industry is headed.

AI-driven decisions are starting to attract scrutiny.
Litigation is emerging.
Regulators are behind, but not indefinitely.

The imbalance is obvious.
Decision velocity is increasing. Oversight is not.

That doesn’t hold for long.

The mistake most payers are making is treating AI like a technology upgrade.

It gets handed to IT.
It gets implemented through a vendor.
It gets measured in terms of cost reduction.

That framing misses the point entirely.

AI in payer operations is not just a technology layer.
It is a decision layer.

And decision layers require control, accountability, and governance.

Right now, many organizations don’t have that foundation in place.

What needs to change is straightforward, but not easy.

Every automated decision needs to be traceable.
Every outcome needs to be explainable.
Every workflow needs to be defensible.

Not in theory. In practice.

Human oversight isn’t going away in high-risk areas.
It just needs to be redesigned around the system, not bolted on after the fact.

AI will continue to expand across payer operations. That’s not the question.

The real divide will be between organizations that deploy it
and organizations that can defend it.

Because the next wave of pressure won’t come from innovation.

It will come from scrutiny.

The question is no longer whether to use AI.

It’s whether your organization can stand behind the decisions it makes when AI is involved.