JustineAI™ — Mass Tort Edition.

1. Stage 01Inventory intake & vetting

   When the MT edition opens, Justine will triage the full plaintiff inventory at acquisition — product exposure, injury type, use dates, prior counsel, lien posture, statute and tolling-agreement status per jurisdiction. Marginal and time-barred claims get flagged before they enter the MDL workup.

2. Stage 02Plaintiff fact sheets

   One supervisor will auto-populate each plaintiff fact sheet from intake records, medical chronologies, and prior productions — every field with per-source provenance. The validator council confirms exposure and injury entries, surfacing cross-source conflicts for attorney resolution before the PFS is attested and served.

3. Stage 03Common-issue discovery

   Justine is designed to read the common-benefit document depository — internal corporate memos, regulatory submissions, adverse-event reports — in one context, surfacing the general-causation and notice evidence that recurs across the pool rather than re-deriving it plaintiff by plaintiff.

4. Stage 04Bellwether selection & scoring

   The bellwether sub-agent will compose PFS density, injury severity, jurisdiction risk, and settlement leverage into a defensible ranking. When CMO 1 sets the trial-selection protocol, the firm walks in with a representativeness argument grounded in the actual inventory, not intuition.

5. Stage 05Science & Daubert

   Justine is designed to assemble the general-causation record — epidemiology, dose-response, mechanism — against each defense expert’s prior testimony and methodological gaps, producing a Daubert pre-read. Every cited authority is grounded against a verifiable source, never invented.

6. Stage 06Bellwether trial workup

   For each selected trial pick, the supervisor will spin a dedicated case workup — complaint refinement, plaintiff-specific causation, deposition designations, motions in limine, exhibit assembly — while keeping every other plaintiff sub-agent in the same reasoning context for cross-matter consistency.

7. Stage 07Global settlement matrix

   Justine will model the inventory settlement as a tiered matrix — injury grids, qualifying criteria, base-and-enhancement point schedules — with per-tier values anchored to public verdicts where available. The matrix exports as a procurement-ready spreadsheet the negotiating committee can defend.

8. Stage 08Allocation & point scoring

   When a settlement is reached, the supervisor is designed to score every plaintiff against the agreed matrix — documenting tier placement, enhancements, and offsets per claimant with provenance — so the special master and claims administrator receive a fully reasoned allocation the firm can stand behind.

9. Stage 09Lien resolution & distribution

   The lien sub-agent will orchestrate Medicare, Medicaid, ERISA, provider, and subrogation liens across the entire pool — computing each plaintiff’s net before distribution, not after. Disbursement statements and an auditable disbursement trail are prepared for every claimant.

• Inventory sub-agent

  Reasons over the entire plaintiff pool at once: product-exposure proof, injury cohorting, use-date windows, jurisdiction and venue, statute and tolling-agreement status, prior-counsel and referral provenance, duplicate and lien-conflict detection across thousands of claimants.

  For example — An inventory of 4,200 acquired claims in a hernia-mesh MDL. The inventory sub-agent will cohort them by implant model and revision-surgery status, flag 137 with no implant record on file, surface 19 whose two-year statute ran before any tolling agreement attached, and identify 6 already represented in a parallel state consolidation — each finding shown with the source page that triggered it, for the intake attorney to clear before workup.

• Bellwether sub-agent

  Reasons over selection-protocol fitness: plaintiff fact-sheet completeness, injury severity, treatment and causation density, jurisdiction risk, defense exposure, demographic and geographic representativeness against the pool, and the CMO’s randomization or strike rules.

  For example — From a 4,200-plaintiff inventory, CMO 1 calls for each side to designate eight trial picks from a representative pool. The bellwether sub-agent will score the inventory, propose a 40-plaintiff candidate set that mirrors the pool on implant model, injury grade, and venue, and explain why its slate is more representative than a naive severity sort — handing the steering committee a Rule-defensible selection argument the special master can scrutinize.

• Settlement-matrix sub-agent

  Reasons over inventory-wide valuation: injury-tier definitions, qualifying medical criteria, base point schedules, enhancement and offset factors, public-verdict comparables, holdback and common-benefit assessments, and the global-deal participation threshold.

  For example — A proposed inventory settlement with a $1.2 billion common fund. The settlement-matrix sub-agent will model a five-tier injury grid, assign base points by surgery status with enhancements for documented complications and offsets for pre-existing conditions, anchor each tier to public bellwether verdicts where available, and project the per-tier average recovery and the firm’s blended fee — exporting it as a spreadsheet the negotiating committee attests before presenting it to the court.

• Lien sub-agent

  Reasons over recovery-side obligations across the pool: Medicare conditional payments and MSA exposure, Medicaid state recovery, ERISA-plan subrogation and Ahlborn reductions, provider balances, and per-plaintiff net-to-client computation before distribution.

  For example — A 4,200-plaintiff settlement with mixed payer profiles. The lien sub-agent will batch-query CMS for conditional-payment ledgers, separate injury-related charges from unrelated treatment, assert procurement-cost reductions on each ERISA plan, and compute the net distribution for every claimant before a single check is cut — flagging 84 plaintiffs whose Medicare exposure exceeds their tier recovery so the attorney can intervene before, not after, disbursement.

• Supervisor over thousands of plaintiff sub-agents

  One Justine coordinates plaintiff fact-sheet drafting, intake triage, and discovery prep across the MDL’s entire plaintiff population — in a single reasoning context.

• Bellwether scoring

  Plaintiff fact-sheet density, injury severity, jurisdiction risk, settlement leverage — composed into a bellwether ranking the firm can defend in CMO 1.

• Settlement-matrix modeling

  Per-tier point-value modeling for inventory settlement. Comparables anchored to public verdicts where available. The matrix exports as a procurement-ready spreadsheet.

• Lien negotiation at scale

  Medicare, Medicaid, ERISA, provider, and subrogation liens orchestrated across the plaintiff pool. The lien total per plaintiff is computed before settlement, not after.

• Plaintiff fact-sheet auto-population

  PFSs drafted from intake records, medical chronologies, and prior productions. Attorney-attested before service.

• Pool-wide triage

  Justine will surface — On a typical inventory acquisition, Justine will read every intake file in one context and surface the claims that don’t belong — the plaintiffs with no proof of product exposure, the statutes that ran before a tolling agreement attached, the duplicates already pending in a parallel consolidation — each tied to the source page, so the firm vets the inventory before it funds the workup.

• Common causation

  Justine will surface — Across a common-benefit depository of internal memos, regulatory filings, and adverse-event reports, Justine is designed to surface the handful of documents that establish corporate notice and general causation for the whole pool — the date the defendant first knew, the buried safety signal — reasoned once and reused across thousands of plaintiff sub-agents rather than re-discovered file by file.

• Bellwether representativeness

  Justine will surface — When CMO 1 sets the trial-selection protocol, Justine will show why a proposed slate of picks does or doesn’t mirror the inventory — flagging that the defense’s eight nominees skew toward the mildest injury grade and the most defense-friendly venues — and propose a representative counter-slate, with the distribution math the steering committee can defend to the special master.

• Settlement-matrix modeling

  Justine will surface — On an inventory deal, Justine is designed to model the injury-tier grid end to end — projecting the per-tier average recovery, the share of the fund consumed by the most severe cohort, and the blended attorney fee after common-benefit holdback — anchored to public bellwether verdicts where they exist, so the negotiating committee sees the consequences of each tier definition before agreeing to it.

• Lien exposure at scale

  Justine will surface — Before distribution, Justine will compute each plaintiff’s net across the entire pool and surface the claimants whose Medicare, ERISA, or provider liens exceed their matrix recovery — the cases that net to zero or below — so the firm renegotiates those liens or revisits tier placement before checks issue, not in a post-distribution scramble.

• Single-context consistency

  Justine will surface — Because the supervisor holds the whole matter in one reasoning context, Justine is designed to catch what fragmented review misses — a causation theory argued for the bellwethers that contradicts a PFS allegation in the broader pool, an expert opinion that helps one tier and undercuts another — keeping thousands of plaintiff sub-agents internally consistent for the attorney to ratify.

The AI reasons; the attorney decides.