Skip to main content

AI Act, MDR/IVDR, QMSR: Practical AI compliance trainings for medical devices

We design on-demand trainings for your roles, maturity level, and objectives — from executive briefings to hands-on workshops. We combine regulatory interpretation and best practice with implementation guidance, grounded in real-world medical device development.

What we offer

Our AI trainings help medical device organizations understand what “good” looks like for AI across the lifecycle — aligned with regulatory expectations and the realities of engineering, data, and product delivery. Each training is custom-built: we clarify your actual goal, identify risks and pain points, and tailor content to your audience — from technical teams to cross-functional stakeholders.

Problems we help you solve

Many organizations face similar questions once AI concepts meet regulatory reality:

  • “Do we have enough data?” — considering intended use, performance, bias, and lifecycle management
  • Evidence beyond performance — traceability, provenance, and controlled development
  • Limits of adaptive learning — safety, stability, and change control implications
  • Supplier control for data providers and governance within quality systems
  • Aligning AI Act, MDR/IVDR, and QMSR requirements
  • Adapting to evolving expectations beyond past audit success

These challenges often highlight gaps between AI development practices and regulatory expectations.

Topics we cover

A. Regulatory landscape & interpretation

  • EU AI Act and its relationship with MDR/IVDR
  • FDA direction for AI/ML-enabled devices
  • IMDRF alignment

B. AI lifecycle expectations

  • Intended purpose, claims, and clinical relevance
  • Data governance: provenance, bias, representativeness
  • Development evidence beyond metrics
  • AI-specific risk management approaches
  • Post-market monitoring and performance drift

C. Quality system integration

  • Integrating AI into QMS processes
  • Supplier and data governance
  • Practical and defensible documentation strategies

D. Standards and references

  • IEC PAS 63621:2026
  • ISO/TS 24971-2
  • Context and applicability of IEC 62304 and IEC 82304-1

Standards are referenced where useful, not as certification targets. Trainings are always tailored to your product, context, and audience.

How it works

  1. Scoping call (30–60 minutes) to align on goals and key questions
  2. Tailored agenda based on your maturity and context
  3. Training delivery (remote, onsite, or hybrid)
  4. Optional follow-up with Q&A or next steps

Format examples

  • Executive briefings (60–120 minutes)
  • Half-day or full-day workshops
  • Multi-session programs
  • Cross-functional alignment sessions
  • Project-adjacent enablement training

What defines our approach

Our AI trainings draw on extensive experience at the intersection of regulation and real-world AI development. We combine regulatory expertise with hands-on technical understanding, ensuring content is both accurate and practical. The focus is on solutions that are defensible, auditable, and sustainable as regulatory expectations evolve.

Get started

Interested in a tailored AI training? Start with a short scoping conversation to align on your goals, audience, and key challenges.

Learning Outcomes

Who it’s for

These trainings are relevant for different roles for different reasons — from strategic decision‑making to hands‑on implementation and evidence generation. They are especially valuable where a shared understanding across disciplines is required.

  • Executive and leadership teams who need decision clarity on feasibility, risk, and investment priorities
  • Product & program owners who translate AI ideas into deliverable plans and evidence
  • Regulatory affairs & clinical teams responsible for claims, intended purpose, and regulatory strategy
  • Quality & process owners who must operationalize controls, documentation, and supplier governance
  • Software/AI engineering & data science teams who build models and need clarity on “what regulators actually expect”
  • Cross-functional groups where alignment is the main blocker (and misunderstandings are expensive)