← Back to the engine

Healthcare · the sector the engine is proven in

AI in healthcare without patient data leaving the hospital.

Validated in Treat Systems' technical environment

The sector the engine was built in, and where it is proven: validated in Treat Systems' technical environment, and in dialogue with a Danish region. Suvard finds CPR numbers, patient names and clinical details in every request and keeps them on the hospital's own network.

The product behind the sector

Healthcare runs on CareProxy

CareProxy is the healthcare edition of the engine: validated in Treat Systems' technical environment, and in dialogue with a Danish region. Danish clinical detection (CPR, patient names, diagnoses), everything on your own network, a signed log where changes can be detected.

Visit careproxy.dk ↗

What Suvard finds here

Danish clinical detection, built for regulated requirements.

CPR numbers
modulus 11 check
Today
Patient and relative names
dictionary + context (Danish clinical language model)
Today
Diagnoses and procedures
ICD-10 / clinical ontology
Today
Medications
ATC
Today
Re-identification from context
combinations of quasi-identifiers
Today

Validated in a Danish health-tech vendor's technical environment and against a purpose-built Danish clinical evaluation corpus. The clinical model is trained on synthetic text only.

The rules it is built for

Supervision: The Danish Data Protection Agency, the Danish Patient Safety Authority and the Danish Health Data Authority (NIS2)

GDPR art. 9

health data (special category)

EU AI Act

high-risk from 2 December 2027 (Annex III, e.g. triage) and 2 August 2028 (CE-marked devices, Annex I)

NIS2

health is an essential sector; logs must be protected against manipulation

MDR

the manufacturer's follow-up on the use of CE-marked devices (Annex XIV Part B)

Suvard provides the technical controls and the evidence; the legal assessment is yours. Not legal advice.

Where clinicians use it

  • Clinical decision support
  • Drafting notes and discharge summaries
  • Documentation from speech (ambient scribes)
  • Answers from the best cloud models, without patient data in the request

Other sectors

And the human in the loop?

It only helps if you can see what the human did.

When experienced radiologists (with over 15 years of experience) were given confident but incorrect AI suggestions in mammography, the rate of correct assessments dropped from 82% to 45.5%. A ''human in the loop'' is not a safeguard in itself. Without an audit trail, no one can subsequently distinguish a deliberate override from a reflex click.

That is why Suvard seals both what the AI delivered and what the human did with it.

Imagine

What if the decision and the outcome were recorded?

Every AI recommendation, what the human did with it, and the final outcome, in one place, independent and verifiable. That gives you more than proof when questions are asked; it gives you the data to see for yourself where AI actually helps and where it does not. You cannot improve a decision that was never recorded.

Next step

A 30-minute check on where AI is already running in your organization.

We start with a mapping, not a sale. You get an honest picture of what passes through your network, and what it would take to be able to prove it afterwards.