Skip to main content
Check a technology
Talk to us
AI RISK PROFILER FOR DEPLOYERS

See AI risk across every product you deploy

Right now you're making deployment decisions on AI products without a consistent way to compare them or understand the potential risks.

ORCHA's AI Risk Profiler, built into Atlas, allows you to see which products carry the highest unmitigated clinical risk, prioritise before deployment, and hold suppliers to a standard that actually means something.

Request a demo
Contact us
Feature 1 _ portfolio (3)

Informed by an international advisory group with representatives from...

APA
MHCC
NHS
APTA1
UNICEF
Allspire
Mind
Ulster
UCL
IBM
DNV
APA Labs
THE PLATFORM

The AI Risk Profiler, built into Atlas

For deployers of AI - like health systems, pharmaceuticals, or regulators - the AI Risk Profiler is accessed through ORCHA Atlas.

In practice, this means: every AI-enabled product in your library gets a structured risk profile, your portfolio dashboard shows comparative risk levels across all of them, and your clinical safety team has a consistent evidence base to work from, rather than a pile of incomparable supplier PDFs.

See the full picture, not a pile of PDFs

One view across every assessed product in your estate. Know in seconds which products need attention before the next procurement sign-off.

Compare suppliers on your terms, not theirs

Every product assessed against the same framework, so you're comparing like with like, not trying to read across five different supplier-defined formats.

Tell suppliers exactly what you need

Set structured AI risk evidence as a condition of procurement. Suppliers know what's expected - and have a clear route to meet it. Procurement conversations move faster, with better evidence on both sides.

Website 540x405 app developer supplier 2
THE CHALLENGE

You're accepting AI risk with every product you sign off - the question is whether you can see it

Every AI-enabled health product you procure introduces specific, identifiable risks - hallucination, misclassification, bias, opacity.

Without a consistent framework, those risks arrive in supplier-defined formats that are impossible to compare, prioritise, or act on.

There's no consistent basis for comparison
No way to prioritise which products carry the highest unmitigated clinical risk
No structured evidence to provide to clinical safety officers or regulators

The AI Risk Profiler makes risks visible

Every product assessed through ORCHA produces the same structured output, comparable across your entire portfolio.

Without AI Risk Profiler

  • Supplier A sends a PDF risk summary with no clinical context

  • Supplier B sends a completed AI Act questionnaire in a different format

  • Supplier C provides no structured AI risk documentation at all

  • Your clinical safety officer has nothing comparable to review

  • You can't prioritise or see which product carries the highest risk

With AI Risk Profiler

  • Every product assessed against the same six-step methodology

  • Risk profiles directly comparable across your whole portfolio

  • Portfolio dashboard: sort by CRS, filter by gaps, flag outliers

  • Clinical safety team has a structured, consistent evidence base

  • Procurement requirements can specify structured AI risk evidence

How the Health AI Risk Profiler works

1

Your health tech suppliers complete the questionnaire

A structured set of questions covering the product's AI features - what they do, how they work, what clinical context they operate in, and what governance and architecture controls are already in place.
2

The tool applies the six-step framework

The tool applies the AI Risk Profile methodology to every AI feature in the product. Each feature is assessed separately - clinical risk severity, computational function, inference technique, which of the 10 standard risks are activated, and what's actually mitigating them.
3

A structured AI risk profile report is generated

A clear, evidence-based output showing which risks are active in the product, how severe each is given the clinical context, where mitigations are working, and where the gaps are.

Designed to support compliance across global AI frameworks - not replace them

Regulatory & Standards Alignment
EU AI ACT
EU AI Act
Articles 9–17 risk management & transparency obligations
European Union
DCB 0129
DCB0129
Clinical risk management for health IT systems
United Kingdom
NICE ESF
NICE ESF
Evidence Standards Framework for digital health technologies
United Kingdom
GMLP
GMLP
Good Machine Learning Practice for medical AI
US · UK · Canada
FDA SaMD
FDA SaMD
AI/ML Software as a Medical Device guidance
United States
ISO 14971
ISO 14971
Medical device risk management standard
International
NHS AI Strategy
NHS AI Strategy
NHS England AI adoption & governance framework
United Kingdom
WHY THIS MATTERS NOW

The legal risk in health AI is already here

The EU AI Act has classified most clinical AI as high-risk, with mandatory documentation obligations live now. In the US, federal regulation is still developing — but the courts aren't waiting.

In the EU, non-compliance with the AI Act's high-risk obligations carries penalties of up to €30 million or 6% of global annual turnover. In the UK, DCB0129 clinical risk management obligations apply to every deploying organisation. In the US, the first cases are already in court. Disclaimers are not a defence.

€30m

maximum EU AI Act penalty

Or 6%

of global turnover

THE METHODOLOGY

You get a complete picture of every AI risk in a product and exactly what to do about it.

Each AI feature is assessed separately. By the end, you know which risks are active, how severe they are, what's mitigating them, and where the gaps are.

Clinical Context — Clinical Risk Severity

The risk profile is anchored in what's actually at stake — who receives the AI output and what happens if it's wrong. A wellness chatbot and a clinical decision tool get very different profiles. That distinction matters, and it's built in from the start.

Computational Function

A product with multiple AI features gets each one assessed separately. You don't get a single blended risk score that obscures where the real problems are - you get a clear picture of each feature individually.

Inference Technique and Model Architecture

How the AI works determines which risks it carries. Suppliers declare this through a structured questionnaire. If the declaration is incomplete or withheld, that's flagged — because opacity in documentation is a risk in itself.

Risk Activation

You find out which of 10 defined AI risks are active in this product — and how serious each one is given the clinical stakes. Where two factors both trigger the same risk, you're told it's amplified. No guessing about what matters most.

Structural and Governance Mitigations

You see what's actually reducing the risk — in the product architecture, and in the governance programme. Assessed separately, so you can tell whether a product is genuinely safe or just well-documented.

Gap Analysis — Residual Risk

You leave knowing exactly where the unmitigated risks are and what they mean clinically. Not a pass or fail — a clear evidence base for making the decision in front of you, whether that's procurement, deployment, or regulatory submission.

THE EVIDENCE BASE

Built on a foundation with AI and digital health experts

The AI Risk Profiler framework has been validated against 20+ AI feature assessments across real health products. Every risk category and mitigation type reflects what's actually in the market - which means when it finds a gap in your product, that finding is grounded in evidence, not theory.

Read more about the research group here.  

30+

Expert institutions in the Advisory Steering Group across clinical, academic and technical backgrounds

20+

Real AI health product features tested during framework development

Pink Glass Icon Search List
RESEARCH FOUNDATION

"Operationalising Trust in Health AI: A Participatory Framework for Contextual Risk Assessment in Digital Health"

A working paper co-authored with Ulster University and the American Psychological Association has been submitted to HHAI2026 - the Hybrid Human-Artificial Intelligence Conference in Brussels, July 2026. A peer-reviewed journal paper is planned for H2 2026.

Learn more

Why this matters now

35%

of people globally now use AI to manage their health, making AI health products a mainstream clinical touchpoint, not a fringe use case.

64%

of AI-fluent consumers believe they can perform at least one medical task as well as a trained professional. Perceived AI competence is outpacing governance.

-10

year-on-year drop in public confidence in finding reliable health information. Trust in the health information environment is deteriorating fast.

Source: 2026 Edelman Trust Barometer Special Report: Trust and Health. 16,009 respondents across 16 markets.

See the AI risk across your technology portfolio

Request a demo to see the AI Risk Profiler portfolio dashboard - risk profiles across multiple products, gap indicators, CRS ratings - and what you could do with that view across your own estate.