Solutions, By function, Medical review

Read it like an MD reviewer

A medical record is eighty to six hundred pages of multi-author free text. The contradictions across notes, the codes that miss the documentation, the absent findings: knowable only if you read it all. Owl reads the whole thing, walks the guideline, and writes a determination an MD can defend.

How a medical record actually flowscohort n=480K files
Records arrive
100%, 80–600 pages each
Baseline
Read in full
28% read end-to-end manually
Industry avg
Cited to span
14% determinations cited to span
Industry avg
IME concordance
64% concord with later IME
cohort study
Owl-determined
92% concord with later IME
cohort study

100%

Records read end-to-end

92%

Concordance with downstream IME

14×

Reviewer throughput

03Reviews by line of business

Pick a line. See how Owl reads the chart, walks the guideline, and writes the determination.

Same three agents, same workflow, every clinical area. The chart and the guideline change; the discipline doesn’t.

OwlVision
What it read
Pre-auth packet + LOMN
18 pp
Read, Indexed
Office visit + imaging history
84 pp
Read, Chronology built
PT records (12 sessions)
42 pp
Read, Conservative-care confirmed
MCG / InterQual criteria
14 pp
Read, Criteria-walked
Structured extraction
Member 54M, L4-L5 lumbar microdiscectomy requested, MRI confirms protrusion + radiculopathy, failed 8 wks PT, 12 wks NSAID, 3 ESI, MCG criteria for lumbar discectomy: documented neurological deficit + failed conservative care, all criteria met
OwlSignal
What it cross-referenced
low
MCG criteria, all elements documented

MCG ortho criteria for lumbar discectomy require: imaging, neurologic deficit, failed conservative care ≥6 wks. All three met with chart citations.

low
Surgeon credentials, board-certified

Surgeon is board-certified ortho spine; hospital is in-network; facility ASC-credentialed for procedure.

OwlAssist
What it answered
Examiner asked
Approve or deny, and on what authority?
Owl answered in 0.6s

Approve. All MCG criteria for lumbar microdiscectomy are met: MRI-documented L4-L5 protrusion with corresponding radiculopathy on PE, 8 weeks of PT documented, 12 weeks of NSAID documented, 3 ESI with partial relief documented. Determination letter cites chart spans and MCG criteria sections. Cohort approval rate on this profile is 96%; appeal-overturn rate <1%. Approval letter and PA-rationale memo attached.

MCG ortho v3.2025MRI 2025-06-14PT records 2025-07Plan §coverage

How Owl plugs into the medical-review bench. Same MDs, more files, every record read.

01OWL

Records land

Owl reads the full chart end-to-end. Builds the chronology, the code map, the contradiction list.

02OWL

Guideline walk

ODG, MTUS, Milliman, MCG, Hayes, NCCN: whichever applies. Cited section by section.

03OWL

Determination drafted

Approve / deny / develop, with chart spans, guideline citations and ERISA-/state-compliant rationale.

04MD REVIEWER

MD reviewer signs

Reviewing MD validates, edits and signs. Owl pre-fills the determination letter and IME questions.

05MD REVIEWER

Appeal defense

On appeal, Owl reads the new submissions and writes the supplemental determination + reviewer rationale.

05What changes

Numbers from carriers running Owl in medical review.

MetricWithout OwlWith Owl
Files read end-to-end28%100%
Determination cycle time4.6 days0.4 days
IME concordance64%92%
Appeal-overturn rate24%8%
Reviewer throughput / MD / day6 files32 files
Medical-spec leakage$840 / file$280 / file
Source: rolling 12-month average across 4 carriers + 2 IROs running Owl in medical review, n = 124K determinations.

Built for the regulatory shape of medical-necessity review.

URAC + NCQAUtilization-management standards: timeliness, criteria, peer-review credentials.
ERISA, §503Reasoned, documented denials with administrative-record support.
ACA, §2719External-review-ready documentation and reviewer credential matching.
State UR + IRO lawState-by-state utilization-review and IRO statutes respected.
HIPAA, HITECHPHI segmented per tenant. Detailed access log per record.
Reg AI / NYDFSBias testing on determination models, explainability, MD-final-decision policy.
SOC 2 Type IIAnnual audit, continuous control monitoring.
MD final decisionOwl drafts; a credentialed MD reviewer always signs the determination.
07Integrations

Lives where your medical review team lives.

UM/UR platforms

GuidingCare, TruCare, ZeOmega, MedHOK, in-house

Claims platforms

Guidewire, Duck Creek, FINEOS, Origami, Majesco

Clinical guidelines

MCG, InterQual, ODG, MTUS, Milliman, Hayes, NCCN

Medical records

MRO, Verisma, Datavant, Epic, Cerner, athena

IRO / IME

External review orgs, IME networks, physician panels

Identity & SSO

Okta, Azure AD, PingFederate, SAML, SCIM

08Get started

Bring us a stack of charts. We’ll send back the determinations your MD bench would have written.

Two-week pilot. 100 records, your guidelines, your MD bench validates against your team’s ground truth. We measure concordance, cycle time, appeal-overturn rate and missed contradictions.