AI does not fail in the model. It fails in the organisation around it.

Hospital and clinic leadership, health authorities, medical faculties and health-technology companies.

Healthcare carries the strictest version of every problem we work on: high-stakes decisions, sensitive data, and a public that will not accept an unexplainable model. The blocking factor is rarely the technology. It is whether anyone has decided who owns the risk.

Services

Four services, with the time each takes

AI maturity assessment

Where the organisation actually sits across informal, functional, integrated and strategic use, and what each level exposes it to. Delivered as a diagnostic leadership can act on, not a score.

4–6 weeks AI governance →

Clinical & administrative AI governance

Risk classification, human-oversight design and documentation for AI in triage, diagnostics and administrative decision-making, built for the deployer obligations rather than the vendor's.

6–10 weeks AI governance →

Health data governance & EHDS readiness

Secondary use, consent architecture and cross-border flows, assessed against GDPR and the European Health Data Space.

6–8 weeks AI governance →

Outcomes & wellbeing measurement

Building measures that hold up analytically, drawing on our work on Nordic wellbeing indices and subjective measurement.

8–12 weeks Data analytics →
Context

What is driving the question here

EU AI Act

Most clinical decision support falls in the high-risk category, and Article 26 puts named obligations on the deployer, not only the vendor.

EHDS

Regulation (EU) 2025/327 changes what secondary use of health data requires, and rewards organisations whose data was governed before they needed it.

MDR overlap

Where a model becomes a medical device, and what that adds to the compliance path.

Informal use

Staff already using free accounts on clinical material is the most common exposure we find, and the one least likely to appear on a risk register.

Engagement formats

Three ways to start

Scoping review

2 weeks · fixed fee

We establish what you are actually deciding, what evidence you already hold, and whether the question needs us at all.

  • Two structured sessions with your team
  • Review of existing documentation and data
  • A written scope with options, effort and sequence

Assessment

6–10 weeks · fixed scope

A structured review against a defined standard — regulatory, methodological or strategic — returned as a prioritised action list rather than a report that sits on a shelf.

  • Baseline against the applicable framework
  • Benchmarking against comparable organisations
  • Findings ranked by exposure, with owners and sequence
  • A workshop handing the findings to the people who act on them

Retained advisory

6–12 months · monthly

For organisations under continuing regulatory or geopolitical pressure, where the question changes faster than a project can answer it.

  • Standing access for named decision-makers
  • Quarterly horizon briefings on your specific exposure
  • Board and committee material on request