Diagnostic and analytical AI
Images, signals, specimens and clinical data become information used for diagnosis or assessment.
Imaging · pathology · laboratoryAI and emerging technologies
Regulate the whole intelligence, not only the algorithm
An AI-enabled medical product is a moving relationship between data, model behaviour, clinical purpose, people and change. We help make that relationship clear enough to evaluate, evidence and control.
informact
fixedadaptive
Dataprovenance
Modelbehaviour
Clinicalconsequence
Changecontrol
AIregulated
system
system
Every position changes the evidence and oversight needed.
A spectrum, not a category
What the technology does matters more than the AI label
Regulatory significance grows with the clinical consequence of the output, the degree of autonomy and the uncertainty introduced by data and change.
Forecasting and decision support
Risk estimates, early warnings and prioritisation influence clinical and workflow decisions.
Triage · prognosis · riskGenerative and conversational AI
Clinical text and recommendations introduce new questions of boundaries, verification and oversight.
Assistants · content · foundation modelsAdaptive and autonomous functions
AI influences therapy, monitoring or device operation with limited human intervention.
Control · automation · personalisationEvidence spine
Performance becomes meaningful only in context
A model metric cannot carry the clinical claim by itself. The evidence must form one traceable line from the origin of the data to the action taken by the user.
Input
Data foundation
Provenance, quality, representativeness, labelling and governance.
Behaviour
Model performance
Independent testing, robustness, subgroup performance and limitations.
Meaning
Clinical relevance
Population, environment, claims, benefit and meaningful endpoints.
Action
Human interaction
Transparency, usability, automation bias, oversight and response.
Regulatory horizon
One product story must cross two systems
The foundation can be shared. Qualification, submissions and lifecycle obligations cannot simply be copied from one market to the other.
Shared product truthPurposeModelDataEvidenceRisk
MDR or IVDR alongside the AI Act
Device qualification and classification, conformity assessment, AI Act role and requirements, technical documentation and post-market duties.
FDA device pathway
Device class and product code, submission route, evidence expectations and a predetermined change control plan where appropriate.
MDRC support
Build a product that can change without losing control
We connect regulatory strategy with model development, software lifecycle, clinical evidence and post-market monitoring. The objective is not only a first submission, but a defensible route for every meaningful update that follows.
Talk to MDRC- 01
Define
Purpose, claims, users, boundaries and route.
- 02
Demonstrate
Data, performance, clinical and human-factors evidence.
- 03
Release
Validated version, transparency and documented controls.
- 04
Learn
Drift, feedback, incidents and real-world performance.
- 05
Change
Planned updates, retraining and regulatory assessment.
