A voice-based hostile cross-examination, generated live from the records you upload. Bates-cited contradiction detection. Six-skill scorecard with specific line-by-line coaching. Built for medical malpractice — used by defense (to prep defendant doctors) and plaintiff (to vet expert witnesses).
Try 2-minute demo Specialty deep-dives →Every question references the records of YOUR case. When the witness says "I don't recall," the AI confronts them with the exact Bates page where they documented the event.
Defendant doctor (defense preps their doctor — hostile plaintiff atty crosses). Plaintiff's expert (plaintiff preps their expert — hostile defense atty crosses with Daubert/FRE 702 attack). Defense expert (defense preps their expert — hostile plaintiff atty crosses on hindsight + bias + scope). Same voice engine, three rubrics.
Volunteered info · Evasion · Contradictions-with-record · Memory failures · Ego responses · Question discipline. Every flag is line-cited with the specific rewrite that would have saved the answer.
AES-256 at rest. Per-case sandbox — no cross-tenant context bleed. Auto-purge in 30 days unless pinned. Anthropic + Deepgram BAA in place. ZDR on every API call.
| Service | Typical price | Format |
|---|---|---|
| Cathy E. Bennett mock-deposition (1 day) | $30,000 | In-person |
| Jury-consultant witness prep | $1,000 / hr | 1:1, live |
| Mock deposition w/ human coach (DepoSim's own benchmark) | $1,800 – $4,200 per session | Live |
| LEAP (Doctors & Litigation) physician coaching | $3,000 – $4,500 for 8–12 hrs | Telephonic |
| Deposition Trainer (this product) | $499 / case · unlimited replays for 30 days | Voice, on-demand |
Carriers typically cap deposition expense coverage at ~$10K. When defense attorney prep + LNC chart organization + a single human-coached mock session blows past that ceiling — which is routine on complex cases — the defendant doctor pays out-of-pocket. $499 covers that overage and leaves change.
— Pricing rationale memo, 2026-05-21Mock-deposition prep is generally protected by attorney work-product privilege, but the existence of a written, time-stamped transcript creates discovery exposure that traditional in-person mock depos don't. We designed four protections in.
When the session ends, the literal Q&A transcript is destroyed and only the summarized scorecard is retained. Opposing counsel asking "what did you review before this deposition" gets back a coaching report, not a Q&A transcript ripe for cross. You can opt-in to keep the full transcript per session if you want it for your own records.
The AI coaches the witness on HOW they answered — volunteered info, ego, multi-part answers — never WHAT they should have said substantively. Substance coaching crosses the witness-coaching line. Our system prompt explicitly forbids it.
For defendant-doctor and fact-witness sessions, the user attests at session creation that an attorney is conducting or directly supervising the prep. This keeps the session inside the work-product privilege envelope — not a standalone consumer service.
When the AI confronts a witness with a record entry, it quotes the chart verbatim. That's surfacing what the existing record already says — not coaching new facts. The chart is discoverable anyway; we're just helping the witness rehearse for the inevitable.
If your jurisdiction has stricter witness-coaching rules (some state bars do — California Rule 3.4 has additional gloss; New York DR 7-102), confer with your local ethics counsel before adopting. The tool's defaults are designed for the ABA Model Rule 3.4 / FRE 612 federal baseline.
The AI examiner is prompted off the canonical medmal cross-exam playbook: Standard-of-care attack · Causation differentials · Hindsight-bias trap · Reptile safety-rule · Ego provocation · Polarizing (Friedman) · Memory-failure exploitation · False-premise / goal-post-move · Prior similar cases / NPDB · COI / financial bias.
The scoring engine is grounded in the SEAK "10 biggest legal mistakes physicians make at deposition," AAFP deposition-defense guidance, and the NITA depo-skills rubric.
The expensive parts of witness prep: human-coached mock depositions typically run $1,800–$4,200 per session, jury-consultant witness prep bills around $1,000/hr, and physician coaching programs run $3,000–$4,500 for 8–12 hours. The Deposition Trainer is $499 per case with unlimited replays for 30 days — a voice-based hostile cross generated live from the records you upload, with a six-skill scorecard after every session.
Mock-deposition prep is generally protected attorney work product, but a written transcript creates FRE 612 exposure traditional in-person mocks don't. Four protections are designed in: purge-on-end is the default (the literal Q&A transcript is destroyed and only the summarized scorecard is retained), coaching is form-only per ABA Model Rule 3.4, defendant-doctor sessions require attorney-supervised attestation, and record confrontations quote the already-discoverable chart verbatim. Confer with local ethics counsel if your jurisdiction has stricter rules.
DepoSim (AltaClaro × Verbit) and ZenWitness ship voice-based AI deposition tools, but neither is medmal-specialized, neither is bar-neutral, and neither cross-references the witness's answers against the actual uploaded chart. This trainer confronts the witness with the exact Bates page where they documented the event — that chart-grounding is the gap it fills.
$499 per defendant per case with a 30-day unlimited-replay window; $1,499 for the three-session prep package (baseline, mid-point, final, with a comparison report); or $199/month for expert prep and vetting (5 sessions per month with a Daubert/FRE 702 reliability rubric). Non-medical expert witnesses can start free at /expert-witness.