Responsible AI implementation

AI works when evidence,
workflow, and review
stay connected.

I help regulated teams turn AI use into documented, source-aware workflows with human control built in.

André Hartmann
45+colleagues supported
22Roche affiliates reached
HTAevidence workflows
QAhuman review controls
AI

The problem

value depends on workflow quality, not demo quality.

AI-supported work can move quickly, but weak source handling and unclear review make outputs hard to trust. Regulated teams need routines that make evidence, assumptions, and responsibility visible.

The question is not whether a tool can draft. The question is whether the workflow can be checked.

That is where implementation work belongs.

The workflow

Practical controls for responsible AI use.

Built from work with evidence, quality, and digital documentation in a regulated pharmaceutical environment.

01 Assess Understand how each team actually works

Map the work before selecting an AI pattern. The important distinction is often whether a team creates new material, adapts existing material, checks sources, or prepares documentation for review. That changes the workflow design and the quality controls needed.

02 Enable Tailor training to each team’s actual tasks

Training works best when it is tied to real documents, real decisions, and realistic review steps. My public CV documents AI implementation support with 45+ colleagues and AI workflow presentations to leads across 22 Roche affiliates.

03 Standardize Document everything so teams function independently

Reusable workflows need simple instructions, decision points, and examples that teams can revisit without a coach in the room. I emphasize process over prompt tricks: what goes in, what must be checked, who reviews it, and what documentation remains.

04 Govern Ensure AI outputs are accurate and compliant

Responsible AI work needs evidence quality checks, citation validation, traceability, and human review. The goal is to make review easier and more explicit, especially in healthcare, life sciences, and public-sector settings where documentation matters.

05 Scale From one team to the entire organization

Scale only after the workflow has a clear owner, documented checks, and realistic training. That keeps AI adoption tied to professional responsibility instead of tool enthusiasm.

What you get

Controls teams can use in real work.

Workflow mapping for regulated teams that need evidence, traceability, and human review

Source and citation validation routines for AI-supported documentation work

Evidence-quality checkpoints for HTA, regulatory, medical, or commercial document workflows

Role-specific training that translates AI use into daily work rather than generic tool demos

Prompt and workflow documentation with versioning, review points, and clear limits

AI tool selection guidance grounded in the task, source material, and review requirement

Who this is for

For regulated teams where evidence and traceability matter.

The useful work is translating AI from a tool into a documented practice: source-aware, reviewable, and clear about professional responsibility.

HealthcareLife sciencesPublic sectorMarket accessMedical affairsEvidence teamsDocumentationAI governance

Why it works

A healthcare and HEOR profile applying AI with professional control.

10+ years

Clinical healthcare across Norway, Australia, and New Zealand.

EU-HEM health economics

European master's training across health economics, HTA, and market access.

3 languages

Norwegian, English, and Spanish for cross-border professional work.

AI use in regulated settings needs documentation, source control, and human review before it scales.

Let's make AI work with the controls your team needs.

Based in Oslo. Open to employers who can engage someone resident in Norway.