JustineAI™ — Insurance Bad Faith Edition.

1. Stage 01Intake & policy analysis

   When the edition opens, Justine will take the matter through structured intake — denial date, line of coverage, claimant’s loss — then read the full policy clause by clause. Coverage triggers, exclusions, conditions, and the reservation-of-rights letter will be cross-referenced against the loss before the file moves.

2. Stage 02Claim-file discovery & review

   Once the carrier produces the adjuster file, Justine is designed to read it end-to-end in one context — claim notes, internal memos, supervisor escalations, reserve changes, and the diary. Nothing is sampled. The handling narrative will assemble itself in chronological order, every entry anchored to its source page.

3. Stage 03Coverage & breach analysis

   Justine will map the carrier’s conduct against its own policy obligations and the implied covenant of good faith. Unreasonable delay, lowball reserve-setting, ignored proof of loss, failure to investigate — each potential breach surfaced with the file entry and the statutory or common-law duty it offends, for the attorney to weigh.

4. Stage 04Pattern-and-practice investigation

   Where the record is public, Justine is designed to scan the carrier’s prior bad-faith litigation, regulatory consent orders, and market-conduct exam findings — surfacing repeated denial scripts, the same claims-handling manual, or the identical exclusion misapplied across matters. Pattern evidence assembled for the institutional-conduct theory, grounded against CourtListener.

5. Stage 05Damages incl. Brandt fees

   Justine will quantify the policy benefits wrongfully withheld, consequential and emotional-distress damages, and prejudgment interest. Brandt fees — the attorney fees recoverable as damages to obtain the withheld benefits — will be computed from actual file effort, with a time-and-rate breakdown the attorney can defend on the stand.

6. Stage 06Claim-handler deposition prep

   Justine is designed to build the adjuster’s deposition outline from three sources: the file’s own contradictions, the handler’s prior testimony where public, and the carrier’s claims manual. Each line of questioning will arrive with the exhibit and the page, sequenced to lock the witness before confronting the entry that impeaches them.

7. Stage 07Punitive-phase preparation

   For the malice, oppression, or fraud showing, Justine will assemble the clear-and-convincing record — managerial ratification, profit-driven denial metrics, the suppressed coverage opinion — alongside the carrier’s net worth from public filings. The pattern evidence from stage four feeds the reprehensibility argument the attorney will present to the jury.

8. Stage 08Trial / settlement

   Justine is designed to model the settlement range against verdict comparables for the venue and the carrier’s documented litigation posture, and to assemble the trial package — exhibit list, the breach timeline as a demonstrative, motions in limine. Every output remains attorney-attested work product; the attorney decides what is filed and what is offered.

• Coverage sub-agent

  Policy-language interpretation. Coverage triggers and exclusions. Conditions precedent. Reservation-of-rights parsing. Proof-of-loss adequacy. Implied-covenant duty mapping. Genuine-dispute doctrine analysis.

  For example — A homeowner’s water-damage claim denied under an anti-concurrent-causation clause. The coverage sub-agent will parse the clause against the cause-of-loss report, flag that the carrier’s own engineer attributed the loss to a covered plumbing failure before the exclusion was invoked, and surface the reservation-of-rights letter sent 47 days after the proof of loss — past the policy’s own acknowledgment window. The genuine-dispute defense will be tested against the file as it stands.

• Claim-file sub-agent

  Adjuster diary reconstruction. Reserve-change tracking. Internal-memo and supervisor-escalation review. Delay-and-denial chronology. Claims-manual deviation detection across the produced file.

  For example — A 4,000-page disability claim file produced in discovery. The claim-file sub-agent will read every diary entry and surface the reserve dropped from $180K to $12K the week after an unsent independent-medical opinion arrived, the supervisor memo instructing the handler to ‘hold pending IME’ for ninety days, and the three benefit letters that quote a definition of ‘total disability’ found nowhere in the issued policy. The narrative will arrive ordered by date, each entry tied to its page.

• Pattern sub-agent

  Prior bad-faith litigation mining. Regulatory consent orders and market-conduct exams. Repeated denial scripts. Shared claims-handling manuals. Institutional pattern-and-practice evidence, public sources only.

  For example — A UIM claim with a familiar denial rationale. The pattern sub-agent is designed to surface eleven prior suits where the same carrier applied the identical ‘minor impact, soft tissue’ framework, a Department of Insurance market-conduct exam citing systematic delay, and a deposition transcript where a regional manager described the very denial metric at issue. Grounded against CourtListener, the matters will assemble into pattern evidence the attorney can offer for the punitive phase.

• Damages sub-agent

  Withheld policy-benefit tally. Consequential and emotional-distress damages. Brandt-fees computation from file effort. Prejudgment interest. Punitive exposure modeling against carrier net worth. Venue verdict comparables.

  For example — Benefits wrongfully withheld total $96K. The damages sub-agent will compute Brandt fees from the hours actually spent recovering those benefits at the venue’s lodestar rate, layer prejudgment interest from the denial date, and model punitive exposure against the carrier’s publicly reported surplus and the reprehensibility factors in the file. Comparable bad-faith verdicts in the county will anchor a defensible range, presented for the attorney’s settlement judgment.

• Policy-language interpretation

  Carrier policies parsed clause-by-clause against the underlying claim. Coverage triggers, exclusions, conditions, and reservation-of-rights letters cross-referenced live.

• Claim-file analysis

  Adjuster file produced in discovery is read end-to-end. Internal memos, supervisor escalations, and reserve changes assembled into a narrative the jury can follow.

• Claim-handler deposition preparation

  The adjuster’s prior testimony, the carrier’s prior settled bad-faith cases (where public), and the file’s own contradictions composed into a deposition outline.

• Brandt fees calculation

  Attorney fees recoverable as Brandt damages computed from the actual file effort, with a defensible time-and-rate breakdown.

• Prior-claim pattern detection

  When public, the carrier’s prior bad-faith litigation and regulatory actions are surfaced as pattern evidence for the punitive-damages phase.

• Policy interpretation

  Justine will surface — Justine will read the issued policy against the denial letter and catch that the carrier quoted an exclusion from a different form edition than the one bound — the operative policy never contained the language the denial relied on. The mismatch will arrive with both form numbers, the declarations page, and the controlling clause, for the attorney to confirm.

• Reserve analysis

  Justine will surface — Across the produced claim file, Justine is designed to surface the reserve quietly cut from $250K to $15K eleven days before the formal denial — while the adjuster’s own diary still recorded the claim as ‘likely covered.’ The contradiction, with both diary entries and the reserve-change log on the same screen, becomes the spine of the bad-faith narrative.

• Delay reconstruction

  Justine will surface — Justine will reconstruct the handling timeline and show that the carrier sat on a complete proof of loss for 214 days, sending four near-identical ‘still investigating’ letters while assigning no new investigative task in the diary. The statutory prompt-payment deadline it blew past will be cited beside the gap, grounded against the governing code section.

• Deposition impeachment

  Justine will surface — Justine is designed to cross-reference the claim handler’s deposition prep against her prior testimony in another bad-faith suit — surfacing that she swore the manual required a coverage opinion before denial in one case and denied any such requirement here. A ready cross-examination sequence will be built from her own contradictory words, each with transcript page and line.

• Pattern detection

  Justine will surface — When the record is public, Justine will surface that the same carrier settled nine prior bad-faith matters on the identical ‘reasonable basis’ theory now asserted, and that a market-conduct exam flagged the practice. Reasoned from real authority on CourtListener, the pattern will be offered as institutional evidence for the punitive phase — never invented.

• Brandt fees

  Justine will surface — Justine is designed to compute recoverable Brandt fees by isolating, from the file’s own time entries, only the effort spent obtaining the wrongfully withheld benefits — not the bad-faith claim itself. The breakdown will arrive at the venue’s lodestar rate with a defensible task-by-task allocation the attorney can present and the carrier will struggle to contest.

The AI reasons; the attorney decides.