The closed-loop PI engine. Detect → CREB™ → CourtChain™ → Legal. Every 2 weeks. Automatic.

One commercial leg. Indifferent to the outcome.

A single recurring fee against total annual dollar volume, charged in marginal brackets -- each tranche of size at its own rate, so the fee always rises with plan size -- billed as an operating expense. No per-bundle charge, no recovery contingency, no percentage of recovery. Your MCO keeps 100% of what it recovers. Industry-standard SIU contingency runs 15-25%; outside counsel charges 33%.

That is not only the easier sale, it is the stronger evidence. A vendor paid a share of recoveries has a financial interest in the finding it is attesting to, and that is the first thing opposing counsel raises when the bundle is offered. A fee that is indifferent to outcome is a materially stronger position in front of a court than any disclosure language.

The fee covers the always-on infrastructure (T1 deterministic scan + Ava + T2 substantiation + T3 CREB™) and the court-ready bundles it produces. Ava is included: clustering, escalation, and continuous watch. There are no per-finding charges, no per-claim charges, no per-FTE add-ons, no per-vertical license expansion.

Retro proves the system. Annual contract compounds it.

Year-1 starts with a one-time retro-active scan ($3.2M, UHG-scale) that builds the base dataset and proves the system on real Snowflake CMS data. The retro surfaces 4 years of accumulated unaddressed leakage in one pass -- that recovery wave realizes over Y1-Y3 as MAC appeals, DOJ FCA filings, and state recovery cycles complete. The annual contract (the flat platform fee) runs in parallel from day one, catching going-forward leakage on nightly cadence.

Neither flow changes what JIL is paid. The fee is the same banded flat fee either way, and it does not move with what you collect. Your MCO keeps 100% of every dollar recovered, retro and going-forward alike.

Pool sizing -- all CMS-public anchors

3-year projection -- Target scenario (25% capture rate)

3-year totals -- all three scenarios

Capture rates anchored to portfolio data. JIL's existing 11-engagement portfolio on real CMS Snowflake data has already surfaced $29.5B at risk with $9.53B recovery high -- a ~32% capture rate against the surfaced pool, sitting between Target and Aggressive bands. The 15% Conservative band assumes only half of currently-surfaced findings clear MAC + appeal + recovery; the 40% Aggressive band assumes JIL Corridor cross-MCO joins are wired in by Y2 to multiply capture. Recovery realization curves (Y1 33%, Y2 40%, Y3 27% for retro; 50% / 75% / 100% for going-forward) reflect typical MAC appeal + DOJ FCA + state MFCU recovery timelines.

What stays the same. What gets force-multiplied.

The cleanest signal that this is not a vendor-displacement pitch: there's nothing in the JIL contract that requires you to change a single existing vendor relationship. Cotiviti continues. HMS continues. Optum-internal continues. Outside counsel continues. JIL runs parallel on a different surface area.

What JIL catches that the current stack architecturally cannot.

These are not vendor failings. They are architectural gaps. The patterns below require multi-vintage joins across 5+ federal datasets running on the entire claim universe -- not on a vendor sample, not on a flagged subset.

Why Cotiviti, HMS, and Optum-internal cannot close this loop.

This isn't a knock on the incumbents. They were built for rules-engine claim editing + sampled audits. They do that work well. What they cannot do -- architecturally, not for lack of effort -- is run the closed loop JIL runs every two weeks. The wedge is real, defensible, and load-bearing on the commercial model.

The one-line summary. Cotiviti gives you a flagged-claim spreadsheet. JIL gives you a court-ready referral packet, every two weeks, anchored on an immutable ledger, auto-submitted to your legal team -- for less money, on a flat fee that takes no share of your recovery. Same SIU. Same vendor stack. Net new recoveries that didn't exist before.