Problem
Prior authorization burden
Cases pile up. Turnaround targets get harder. Clinical reviewers spend time on routine work and hard exceptions alike.
Prior Authorization
Review burden is the economic problem in prior authorization.
Prior authorization already costs clinical time, turnaround pressure, and exception work. AI can remove routine steps. It can also grow full review if the plan cannot tell a well-supported case from a risky one.
Start with volume, full-review share, time, and loaded cost. Then talk about review burden, economics, AI maturity, and decision control.
Start with review burden.Then economics, AI maturity, and Decision Control.
The problem
PA work is heavy before AI. AI can make the economics worse if full review grows with volume.
Intake, extraction, criteria, clinical review, and downstream notices already stack. Automation helps only when full review, exceptions, and rework do not take the return.
Cost
Review and exception cost
Every full review has loaded cost. Avoidable full reviews erase automation savings before the model score looks wrong.
Better state
Right review at the right time
Supported work moves with less friction. Hard cases get clinical attention. Leaders can explain why a payer action was allowed.
When review is required, it should be substantive: the reviewer receives enough relevant context and real decision authority to exercise independent judgment before the consequential action moves forward.
Review economics
Use your volume. See what a lower full-review share could mean.
Enter annual authorization volume, the share that gets full review today, average review time, and loaded reviewer cost. Scenarios show 5%, 10%, 20%, and 30% reductions in full reviews. These are economic planning scenarios, not realized customer outcomes.
Review-burden model
These are economic planning scenarios, not realized customer outcomes.
Related payer examples
The broader payer lesson
These are related payer examples about review quality and practical AI authority. They are not presented here as prior-authorization case facts.
Substantive review
Cigna / PxDx reporting
What happened. ProPublica reported rapid claim denials with limited clinician time per case under Cigna's PxDx system. Cigna has stated that physicians review claims and disputed the characterization.
Why it matters. Named human review is not the same as substantive review with enough context and real decision authority.
Where Decision Control could intervene. This is a point where Decision Control is designed to intervene when review is named but may not be substantive for the case at hand.
What different handling could have occurred. Cases that look routine on paper could have been routed to qualified review with enough context and authority before a denial posted.
Practical authority
UnitedHealth / nH Predict litigation
What happened. Plaintiffs allege nH Predict shaped post-acute care cutoffs over treating clinicians. The company states clinicians make medical necessity determinations and the tool guides care.
Why it matters. Leaders need a case-appropriate answer for authority, verification, and qualified judgment before the payer action moves.
Where Decision Control could intervene. This is a point where Decision Control is designed to intervene when an AI-shaped recommendation begins to carry practical authority over care.
What different handling could have occurred. The case could have required verification, been routed to qualified review, or been held or escalated before the action posted.
For prior authorization, the operating questions remain: which cases can proceed with less friction, which need verification, and when required review is substantive enough to change the outcome.
CMS context
CMS-0057-F timing is already an operating priority for impacted payers.
Process rules generally began January 1, 2026. API rules generally begin January 1, 2027. Exact dates vary by payer type. This is timing context, not legal advice.
2026 process
Decision timeframes, denial reasons, and metrics
Impacted payers generally must meet process rules starting January 1, 2026, including decision timeframes for most payer types, specific denial reasons, and public metrics (initial metrics by March 31, 2026 per CMS).
2027 APIs
Patient Access, Provider Access, Payer-to-Payer, Prior Authorization APIs
Major API work is generally due beginning January 1, 2027. Timing varies by payer type.
Operating bridge
Electronic PA is not the same as operating value
Data exchange helps. Plans still need the right review posture and a clear basis for consequential actions.
Decision Control
Proposed application: Decision Control for prior authorization
When AI influence builds across intake, extraction, criteria, and clinical review before a payer action, Decision Control applies selective control: more AI work where the case supports it, tighter human handling where consequence and uncertainty rise. It helps set case-level responsibility, direct review when needed, protect automation value, and keep a clear record.
Case-to-case responsibility
AI authority can change with the case, evidence, and consequence.
Direct review when needed
Send the right cases for extra clinical review instead of reviewing everything the same way.
Protect automation value
Keep human handling where it is structurally required, not only where current controls cannot discriminate.
Keep a clear record
Preserve a basis for why a consequential payer action was allowed.
Public handling language: Proceed, Verify, Review, Hold or escalate. Explore Decision Control Plane
Next step
Evaluate the workflow before AI takes on more PA responsibility.
Bring one prior-auth workflow. We will recommend Initiative Review, Build, Pilot, Decision Control Assessment, Decision Control Plane, or no engagement.
AI Initiative Review
$3,500 · 5 business days.
See Initiative ReviewPilot Efficacy & Scale Readiness
Scoped. Prove whether the pilot should scale.
See Pilot EvaluationDecision Control Assessment / DCP
DCA starts at $30,000. DCP is scoped.
See DCAEvaluate the workflow
Bring review burden, economics, and the Decision Control question. Start with a Decision Control Assessment when AI already shapes consequential PA actions.