I have heard that sentence in almost every meeting I have taken for the last six years. It is usually said politely, near the end, and it is almost always true.
The plan does have a vendor for that. It has a vendor for chart retrieval and a vendor for coding. It has one for quality reporting and another for submissions. It has analytics on top of all of it and a consultant who helps interpret the analytics. Every function in the operation is covered by someone competent who can produce a report proving they did their part.
And yet when I ask a straightforward question about a single member, what conditions are real, what care was delivered, what evidence supports it, the answer takes days and requires four people. That gap is what I want to talk about, because I do not think it is anyone’s fault. I think it is the natural result of how our industry got built.
Nobody chose this architecture. It accumulated.
Every category in this market was created by a vendor who solved one visible failure. Somebody built a better way to retrieve charts. Somebody built a better way to find missed conditions. Somebody built a better way to get files accepted. Each one was a real improvement on a real problem, and each one sold well for exactly that reason.
Health plans bought the same way. A submission deadline goes badly, so the plan buys a submission solution. An audit produces an uncomfortable finding, so the plan buys audit support. A quality score drops, so the plan buys a quality vendor. Every purchase was a reasonable response to something that had just happened.
What no plan ever did was sit down and design how information should move through the organization. That was never on anyone’s agenda, because there was never a month when it was the most urgent thing. So the architecture is the accumulated residue of a decade of sensible individual decisions, and nobody owns it.
Every function is covered. The member is not.
A plan gets paid for the risk it can prove, judged on the quality it can show, and audited on the evidence it can produce.
Those are three different scorecards, and the plan usually treats them as three separate programs with three separate budgets and three separate vendors. But they are not three things. They are one member, one clinical record, and one set of evidence, being asked three different questions by three parties who never see each other’s work.
This is why the answers disagree. Not because anyone is careless, but because no one is looking at the same picture. The plan ends up with excellent coverage of every function and no coverage at all of the thing the functions exist to serve.
Connection is not intelligence. We spent twenty years connecting everything and still cannot act on the full story of one member.
The question is not who covers the function
When a plan tells me they already have a vendor for something, they are answering a question about coverage. That was the right question fifteen years ago, when the problem was that these functions were being done manually or not at all.
The question now is whether anyone in the organization can see the member whole, in time to act. Not in a report produced next quarter that explains what went wrong. In time to do something about it while the year is still open.
That is a different kind of capability, and it is not purchasable one function at a time. You cannot assemble a view of the whole from parts that were each designed to answer a narrower question and discard whatever they did not need.
What an operating layer actually does
What payers need is an intelligent operating layer. One version of the truth underneath everything else, holding the member, the evidence, the money, and the risk in one view.
In practice that means the clinical record is read once and completely, and everything downstream draws from the same reading. The condition, the evidence behind it, the quality measure it touches, what was submitted, and what came back are all attached to the same member in the same place. When someone asks why two numbers disagree, the answer is available rather than assembled.
Our AI surfaces what matters and never invents a diagnosis. A certified coder reviews the evidence before anything reaches the government. We call that coder-in-the-loop, and it is the opposite of the black box AI most of this market is selling.
The goal is not to remove the people who understand this work.
The goal is to give them the same evidence at the same time.
What this changes for the people doing the work
A coder stops working from a chart that arrived without context and stops re-deriving what someone else already established. An analyst stops spending the first three days of every investigation assembling a picture before the actual work can start. A leader stops discovering in March what was knowable in October.
None of that is glamorous and none of it makes a good demo. It is also the entire difference between an organization that reacts to what already happened and one that can act while it still matters.
Why this is urgent now
Across Medicare Advantage, the exchanges, and Medicaid, plans are facing rising costs, tougher audits, and shifting risk models at the same time. Some are cutting benefits. Some are leaving markets entirely.
Not because they stopped caring about their members, but because their operating model was never built for the intelligence that survival now demands. That is the part I find worth fighting about. The decision to walk away from a population is often made by leaders who cannot see clearly enough to make a different one.
The next decade will belong to the organizations that turn disconnected information into operating intelligence. Not by replacing people, but by giving coders, clinicians, providers, and payer leaders the same evidence at the same time, before the system fails the member.
So the next time someone tells me they already have a vendor for that, I am not going to argue. They do. I am going to ask a different question instead. Can anyone here see this member whole, today, while there is still time to do something about it?
See what one version of the truth looks like across your own operation.
Frequently Asked Questions
What is payer intelligence?
The ability to see a member completely and act on it in time. It covers what conditions are real, what care was delivered, what evidence supports it, what was paid, and what an auditor would find. Payer intelligence treats those as one picture rather than as separate reporting programs.
How is that different from analytics?
Analytics explains what already happened, usually after the period closed. An operating layer holds one version of the truth underneath the work itself, so the information is available while decisions are still open rather than described afterward in a report.
Does this require replacing every vendor a plan has?
No. The point is not consolidation for its own sake. It is that one complete view has to exist somewhere. Plans can reach that in stages, and the practical starting place is usually wherever the disagreements between numbers are most expensive.
Why can most plans not answer a simple question about one member?
Because the information is spread across systems that were bought at different times to solve different problems, and none of them was designed to hand off a complete picture. Each one kept what it needed and discarded the rest.
What does coder-in-the-loop mean?
Our AI surfaces evidence and recommends, and a certified coder reviews that evidence before anything is submitted. No diagnosis is generated automatically. It is the opposite of a black box that produces an output nobody can explain when an auditor asks.
Does this apply outside Medicare Advantage?
Yes. The same structural problem exists in the exchanges and in Medicaid, with different rules and different models on top of it. Plans running multiple programs usually feel it most, because the same member data is being asked different questions by different teams.
What should a leader look at first?
Pick one member whose numbers disagree across programs and ask the organization to produce the full picture. How long that takes and how many people it requires is a direct measurement of what the organization can actually see.
Is this a technology problem or an operating problem?
It is an operating problem that technology can finally address. The tools to read a record completely and hold it in one place now exist. What has been missing is a decision to treat the member view as infrastructure rather than as the output of whichever program asked last.
