How Clinical Intelligence Eliminates Health Plan Friction
-
Click here to listen to this article - Share via
- AHIP and the Blue Cross Blue Shield Association reported in April 2026 that member plans had cut prior authorization volume by 11%, roughly 6.5 million fewer requests.
- A 2026 review in npj Digital Medicine documents rising AI-assisted coverage denials in Medicare Advantage, and calls for more transparency in how these tools approve or deny claims.
- Utilization management and payment integrity teams typically operate on separate systems, so providers submit the same patient data more than once.
The Shift From Administrative to Clinical Measurement
AHIP and the Blue Cross Blue Shield Association reported in April 2026 that participating health plans had reduced prior authorization volume by 11%, roughly 6.5 million fewer requests. The standard for success is now moving. Under the Centers for Medicare and Medicaid Services contract year 2027 final rule, Medicare Advantage and Part D plans face CMS Star Ratings measurement weighted toward clinical care, outcomes, and patient experience rather than administrative process.
Health plans are buying artificial intelligence to meet those expectations. Vendors selling into that market increasingly pitch connected clinical intelligence, linking systems across departments rather than adding them one function at a time. Bolting isolated tools onto already fragmented workflows adds complexity, and the resulting tech stack creates friction across utilization management, appeals, and quality reporting.
Matt Parker is Chief Product Officer at Cohere Health. He spoke with LA Times Studios in August 2026 about how it works in practice, what the plans using it report, and what payers will be measured on next.
Cohere Health puts its reach at 25 million covered lives across 17 client health plans, and 47 million plan and provider interactions a year. Its materials describe the division of labor as artificial intelligence handling administrative work while humans stay in control of critical decisions [2].
The field is also under scrutiny. A 2026 review in npj Digital Medicine documents a rise in AI-assisted blanket coverage denials, concentrated in Medicare Advantage plans, along with the class action lawsuits and congressional investigations that followed [1]. Its authors call for more transparency in how these tools approve or deny claims, and for attention to the provider burnout and patient burden created by navigating prior authorization requests and appeals.
The Operational Gap in Utilization Management
Kevin Famuyiro: Health plans frequently add separate software tools for individual tasks, which increases system complexity. What has to be true, technically, for a plan to share unstructured clinical data across utilization management, appeals, and payment integrity at once?
Matt Parker, Chief Product Officer at Cohere Health: Health plans often use different software systems that don’t talk to each other, so the same patient information gets requested and managed separately by different teams, even when it’s needed for related decisions. That’s a missed opportunity: connecting insights across teams (not just within them) is where the biggest gains in speed, cost, and quality of care happen. For example, notes from a doctor’s visit can also help speed up payments or flag patients who need preventive care.
We built Cohere Unify, a clinical intelligence platform designed to connect the dots across a health plan’s operations. Cohere Health’s 70 physicians work with our data scientists to train our AI on real clinical evidence and decisions (not general-purpose AI like the chatbots most people are familiar with) to turn messy, unstructured medical records into clear insights that support care decisions for hundreds of millions of people.
By linking policies, care management, payments, and more, Cohere Health is helping build a smarter healthcare system, one where every decision informs the next, making access to care faster and more affordable.
Kevin Famuyiro: Utilization management teams and payment integrity units historically review the same clinical data at different times. What operational friction occurs when clinical details from initial care authorizations fail to reach downstream claims audits?
Matt Parker: Today, care planning (utilization) and claims review (payment integrity) run on completely different teams and systems. Providers share patient data, diagnoses, tests, charts, during planning, then often have to share it all again when the bill comes in.
We collect that data upfront, before care happens, and share across the organization, so everything downstream runs more smoothly. This cuts providers’ paperwork, gives health plans more consistent information, and helps prevent claims denials, so providers get paid right the first time.
Kevin Famuyiro: Cohere Health says one of its pre-payment audits caught a mismatch between an authorized pneumonia diagnosis and a billed sepsis diagnosis, recovering over $5,000 on that claim. How does Cohere Unify analyze unstructured clinical records to flag that kind of discrepancy before claims are paid?
Matt Parker: Consider this scenario: a patient is approved for pneumonia treatment, but complications lead to sepsis, a serious organ-failure condition faced by close to 2 million people each year. When the care needs shift, so can the billing, which can mean higher costs and more work for providers and health plans, plus avoidable costs for patients.
Since hospital billing teams work separately from care teams, tracking down the right data traditionally means digging through charts, reconciling notes, and exchanging countless faxes and calls.
Cohere Health uses AI to read medical records, summarize care, apply payment rules and contracted rates, and reconcile the different clinical scoring systems and medical guidelines (like SOFA scores for organ function) that organizations use. This cuts administrative burden, helps physicians get paid faster and more accurately, and streamlines billing for health plans and hospitals alike, ultimately improving both healthcare and the patient experience.
Automating Quality and Care Exclusions
Kevin Famuyiro: Measuring the effectiveness and quality of care is important in the healthcare system. What factors make that challenging for health plans and prevent them from helping their members get care?
Matt Parker: I agree that the effectiveness and quality of care are non-negotiable. Frameworks like HEDIS, the Healthcare Effectiveness Data and Information Set, and CMS Star Ratings for Medicare Advantage give health plans a common language for quality. One measure example is whether patients get appropriate treatments and preventive services.
This relies on medical records and claims data held in different places within a health plan, and even outside of their organizations with physicians.
Today, health plan nurses manually review patient medical charts one by one to identify and close care gaps, a costly, resource-intensive process that pulls these skilled nurses away from valuable clinical work. By using data from care planning and other sources, our platform identified thousands of instances where patients didn’t get preventive care; this was done without human involvement.
Those nurses then focused on helping people get the care they need, and spent less time on manual reviews. Ultimately, it means improving patient health for people around the country.
Kevin Famuyiro: Cohere Health reports that connecting authorization records to quality systems surfaced 60% of frailty exclusions and 30% of cardiovascular exclusions that manual reviews missed. How does the software distinguish between a care gap and a clinically valid medical exclusion?
Matt Parker: Some patients have legitimate medical reasons to be excluded from a given healthcare quality measure or score, for example, a frail or seriously ill patient might not be the right fit for a treatment, so they should be excluded. Missing those exclusions makes a health plan’s quality score look worse than it should, and it’s costly if teams manage unnecessary care.
Cohere Health’s technology, which is built specifically for HEDIS measures, catches those exclusions by connecting clinical data from utilization management, gathered when care is first authorized, to quality reporting. For one health plan, that connection uncovered 60% of frailty exclusions and 30% of cardiovascular/advanced-illness exclusions that would otherwise have gone undetected, meaning people are getting appropriate care, too.
Connected Clinical Intelligence and the Next Performance Measures
Kevin Famuyiro: AHIP and the Blue Cross Blue Shield Association report an 11% drop in prior authorization volume across participating plans, roughly 6.5 million fewer requests. How does sharing real-time clinical context between plans and healthcare providers change daily clinical workflows for treating physicians?
Matt Parker: Yes, several leading health plans recently announced their success in cutting administrative red tape. At Cohere, we’re helping health plans around the country streamline access to care in three ways.
First, we help them identify treatment requests that don’t need authorization, which helps reduce the volume you mentioned.
Second, we help physicians and their teams reduce their administrative burden with clinically trained AI and automation, which also gets patients to appropriate care more quickly. Health plans using Cohere Health’s technology achieve 94% provider satisfaction, and those providers are seeing faster access to care for their patients.
Our deep clinical intelligence connects symptoms, treatments, diagnoses, policies, and evidence-based medical guidelines so that care can be approved in seconds. This means patients can visit the doctor and have their next steps for treatment approved before they leave the office.
Third, we’re using the data from this process to help health plans connect people with needed preventive care, and to get providers paid more quickly and accurately.
Kevin Famuyiro: As health plans adapt to regulatory changes over the next 1 to 3 years, what specific performance metrics will distinguish organizations using unified clinical data from those operating separate department databases?
Matt Parker: Today, health plans are using technology to optimize operations within functions. For example, physician caseload turnaround times, care gaps closed, and similar efficiency measures.
These are important, but they’re still optimizing within a specific team.
By eliminating operational fragmentation that exists today and using connected clinical intelligence and unified data, health plans can identify opportunities that span departments and provide a holistic view of a patient’s health. For example, tracking metrics like overall patient wellness, preventative care rates, and timely follow-up after discharge from a hospital quickly and more accurately.
At Cohere Health, we see even bigger opportunities to make the entire healthcare system more efficient and effective. Where every function generates a signal for new performance transformation. And where teams like utilization, claims, payment integrity, care management, quality, and risk adjustment operate as one.
They can talk about performance outcomes in economic terms, like: we’re reducing cycle time on our highest-volume care category from 9 days to under 24 hours. Leading healthcare organizations can make healthcare measurably more efficient and affordable.
References
1. Raza, S., Gerke, S., Silcox, C., Hendricks-Sturrup, R., & Shachar, C. (2026). Medicare advantage becoming a disadvantage with use of artificial intelligence in prior authorization review. npj Digital Medicine, 9(1). https://doi.org/10.1038/s41746-026-02387-x
2. Cohere Health. (2026). Clinical intelligence that powers health plan operations. https://www.coherehealth.com/
