Services

Six ways to bring clinical rigor to your AI.

Each engagement is scoped to your product surface and risk profile. Most begin with a discovery call and a short written proposal.

Methodology

Reviews typically involve structured interaction testing, failure-mode analysis, trajectory review across multi-turn conversations, and evaluation against established clinical standards of care. Findings are written in language engineers and clinicians can both act on.

01

Clinical Red Team Review

Adversarial probing of model behavior across therapeutically loaded scenarios, run by a clinician rather than a generic safety tester. Severity-ranked findings with remediation guidance.

Typical fit

  • ·Pre-launch readiness
  • ·Vendor due diligence
  • ·Quarterly clinical review
02

Mental Health Safety Audit

Targeted assessment of how the system handles suicide ideation, self-harm, abuse disclosure, psychosis, intoxication, and other high-acuity moments. Mapped against established standards of care.

Typical fit

  • ·Crisis flow audits
  • ·Escalation policy design
  • ·Incident retrospectives
03

Therapeutic UX Evaluation

A clinician's read of language, pacing, register, and the felt experience across long conversations. Where the product builds — or quietly erodes — the user's sense of being met.

Typical fit

  • ·Voice & tone systems
  • ·Conversation design review
  • ·Onboarding & rupture flows
04

Prompt & Conversation Review

Line-by-line review of system prompts, persona definitions, and multi-turn transcripts. Identifies where instructions invite sycophancy, premature interpretation, or scope drift.

Typical fit

  • ·System-prompt iteration
  • ·Persona design
  • ·Transcript QA pipelines
05

Psychometric Evaluation Design

Design of measurement instruments, rubrics, and rater protocols for evaluating model behavior — drawing on a century of psychological measurement rather than ad-hoc benchmarks.

Typical fit

  • ·Eval set construction
  • ·Rater training
  • ·Outcome instrumentation
06

Advisory Retainer

Ongoing clinical counsel for product, research, and policy teams. Embedded review of new features, evals, datasets, and externally-facing claims.

Typical fit

  • ·Founding clinical advisor
  • ·IRB / ethics input
  • ·Regulatory positioning

How engagements run.

Transparent, scoped, and designed to leave your team more capable than when we arrived.

  1. 01

    Discovery

    A 45-minute conversation to understand the product, users, and the questions you most want answered.

  2. 02

    Proposal

    A short written scope with timelines, deliverables, and a flat fee. No retainers required.

  3. 03

    Evaluation

    Hands-on review of the system, supported by structured clinical protocols and conversation probes.

  4. 04

    Findings

    A written report, a working session with your team, and prioritized recommendations you can act on.

Tell us what you're building.

Most teams reach out before launch, but useful work can begin at any stage. Early conversations are free.

Request a consult