For law firms using generative AI · 5-day fixed scope

Legal AI Practice Audit — The Firm-Level Answer to the Questions Coming

Bar-ethics opinions are moving quarterly. Sophisticated clients are asking, in outside-counsel guidelines, how you use AI on their matters. E&O carriers are updating malpractice questionnaires every renewal. The tools you use don't answer these questions. This audit does.

$2,500 USD · boutique tier · 5 days
Start the audit What's included

Independent by design

Every vendor whose tool you use has a story about their reproducibility posture. Some of the stories are good. Most are marketing. A firm relying on the vendor's own account is in the same trap as a company relying on its own AI's self-grading — the vendor cannot, structurally, produce the third-party attestation your clients, carriers, or bar committee want. Same reason firms hire outside counsel to review their own conflicts practices: independence is the value.

What you get

Four deliverables. Five business days from data handoff.

The audit

  • AI-use inventory — every AI tool the firm uses, mapped to every practice area and matter type
  • Reproducibility posture per tool — model versioning, output retention, log-of-record, cache determinism, tool-trace availability
  • Bar-ethics compliance map — state by state where the firm practices, what each requires, and where the firm satisfies or gaps
  • Client-facing defensibility statement — drop-in language for outside-counsel-guideline responses + client engagement letters
  • E&O carrier answer sheet — drop-in for annual malpractice questionnaires
  • 30 / 60 / 90-day roadmap — specific closes for identified gaps, ranked by exposure

You provide

  • List of AI tools the firm currently uses (best guess is fine; we'll surface the gaps)
  • Practice areas + matter types AI touches
  • The specific client questionnaire, carrier question, or bar-ethics concern that triggered the engagement (if any)
  • 30-minute kickoff call w/ managing partner OR firm risk officer OR CIO

The three surfaces where firm AI-use goes wrong

Cross-tool drift
Most firms use 3-5 AI tools (Harvey, Casetext, Aracor, Ivo, Callidus, Wordsmith, etc.). Each has its own versioning, log format, and retention posture. When you need to produce "the AI record" for a matter, it's fragmented across vendors that don't coordinate.
Silent version drift across matters
Harvey in March is a different model than Harvey in September. The UI didn't change. No lawyer noticed. Client A's matter and Client B's matter, side-by-side in discovery, show different-shape AI reasoning that the firm can't easily reconstruct.
Confident-but-wrong that survives lawyer review
Not the fake case citation (lawyer catches most of those). The subtle clause misinterpretation, the systematic discovery mis-classification, the redline that flags the wrong pattern. Looks like a judgment call. Isn't.

Pricing tiers

Fixed-scope, fixed-price. No hourly billing.

Solo / Small
$500-$1,500
Single-tool audit · 2 days · 2-page brief + defensibility statement
Boutique (5-30 lawyers)
$2,500
Full 4-deliverable audit · 5 days · walkthrough call included
Large firm (per practice area)
$10K-$25K
Multi-tool + multi-state + partner-attestation prep · 2-4 weeks

Start the audit

Boutique tier — $2,500 flat, five business days from kickoff. If your firm has an active client or carrier question on AI-use, say so on the kickoff call — I'll compress the timeline if possible.

Buy the audit — $2,500

Prefer a 15-minute qualification call first? Email luddy.kevin@gmail.com.

Why me

Kevin Luddy. 35 years shipping software. 24-year CTO at Oculearn (HIPAA-adjacent clinical platform). Not a lawyer — a technologist auditing legal-tech.

The audit surface is technical (model versioning, output retention, log-of-record, cache determinism, tool traces) crossed w/ regulatory (state ethics opinions, E&O carrier language, sophisticated-client contract terms). Neither pure lawyer nor pure engineer produces a defensible cross-disciplinary artifact. contrarianAI is the position that does — independent, technical, and calibrated to the specific questions your firm's counterparties will ask.