A claim about another claim
A stop-loss claim is a claim about another claim: one catastrophic medical expense, with disclosure, lasering and TPA-supplied detail behind it. Owl reads the medical record, the disclosure, the eligibility file and the TPA pay file at once, walks the contract, and tells you whether it breaches.
$1.45M
Cumulative paid, day 226
$1.0M spec
Specific deductible attachment
$450K
Stop-loss recoverable day-one
Three AI agents on a stop-loss file. Reading. Cross-referencing. Answering.
One claimant: $1.45M paid, day 226, recoverability answered in 1.6s.
Anonymized composite. Self-insured employer, $1.0M spec, $1.5M agg corridor, claimant 6F with relapsed leukemia and CAR-T therapy.
External data only. None of this lives in your CMS. Owl gathered it from public records, professional networks, registries and the open web, then resolved it to this claimant.
Owl’s CAR-T cohort cost-trajectory model (n=4,200 claimants across self-insured plans) places this claimant at the 56th percentile for cumulative paid by month 7 post-CAR-T. Within reasonable band; no TPA mispayment pattern detected.
Disclosure named "leukemia, in remission, last treatment 2022" with no expected high-cost therapies. Chart shows surveillance imaging in 2023 with rising MRD that was not disclosed. Material? Likely no, because the policy is post-bind and the relapse occurred after policy effective. But flag for underwriter for laser-renewal decision.
Policy exclusions checked: experimental/investigational (CAR-T is FDA-approved for this indication, exclusion does not fire), congenital (does not fire), TPA-error (no error pattern), eligibility (member fully eligible).
Treating institution is an NCI-designated comprehensive cancer center, in-network with the underlying medical plan. No center-quality or unbundling pattern detected.
Yes, recoverable. The claimant breached the $1.0M specific deductible on day 198 of the policy year. Cumulative paid is $1.45M as of day 226. Recoverable today is $450K (cumulative paid above spec). The aggregate corridor is not in play; total plan-paid sits below 125% of expected. No exclusion fires (CAR-T is FDA-approved for relapsed B-ALL), no eligibility defect, and the disclosure-versus-chart comparison shows MRD surveillance not disclosed, but the relapse occurred after policy effective, so it is non-material to coverage. Pay the claim and route the disclosure note to renewal underwriting for the laser conversation.
Lean yes, but laser at $250K, not exclude. Relapse + CAR-T claimants run a 38% second-relapse rate within 24 months per the cohort trajectory model, with average second-relapse spend of $620K. A $250K laser captures expected continuing surveillance and any second-relapse spend up to that band, while keeping the placement competitive. Memo drafted with cohort tables and prior-treatment timeline.
What the world says about your Stop-loss claimants, none of which lives in your CMS.
Internal claim data is half the picture: it tells you what the claimant told you. OwlSignal continuously gathers external data (public records, professional and corporate registries, court dockets, social and open web) and matches it back to the claimant. These are the families that move the needle on Stop-loss.
Owl’s cross-payer cohort cost-trajectory models for the catastrophic conditions that drive stop-loss exposure: CAR-T, BMT, ECMO, transplant, gene therapy, hemophilia, premature neonate.
Disclosure statements compared element-by-element to the underlying chart. Eligibility files, COBRA records, and certificate-of-coverage tested against the claimant’s active dates.
Cross-carrier prior medical history, where law and consent allow, to surface prior conditions, prior procedures and prior failed therapies relevant to material-disclosure analysis.
NCI-designated centers, transplant program registries, OIG LEIE, state medical board sanctions, and Owl’s cross-payer outcomes graph for catastrophic-care providers.
Stop-loss policy, endorsements, lasers, exclusions, run-in / run-out, eligibility waivers: read once, walked on every breach decision.
TPA paid-claim ledger re-tested against policy: timely-pay, FFS coding, unbundling, balance-bill exposure, claim age. The integrity check that decides what is and isn’t reimbursable.
Numbers from MGUs and reinsurers running Owl on stop-loss in production.
Built for the regulatory shape of Stop-loss.
Lives where your Stop-loss files live.
Stop-loss platforms
Sun Life, Tokio Marine HCC, Symetra, QBE, in-house
TPA platforms
WLT, Trizetto QicLink, LuminX, Plexis, in-house
EHR / records
Epic, Cerner, MRO, Verisma, Datavant
Disclosure intake
Box, OnBase, ImageRight, email parser
Cohort + cost references
CMS DRG, NCCN cost, Milliman MARA
Identity & SSO
Okta, Azure AD, PingFederate, SAML, SCIM
Bring us a quarter of large claims. We’ll show you the breach math on day one.
Two-week pilot. Your large-claim stream, your tenant, your policy. We measure cycle time, leakage capture, disclosure-defect detection and laser-renewal support against your own ground truth.