← Back to the engine

Legal · later

AI in law firms without a client's case leaving the office.

Detector not built yet

A lawyer's duty of confidentiality is broader than GDPR: it covers everything about the client's case, not only personal data. A name, a case number or one sentence about what a client is considering must never reach a cloud model. The engine is the same as in healthcare; the legal detector is not built yet.

What Suvard would protect here

The entities fit the engine. Confidentiality without clear markers requires case-scoped routing.

Client and counterparty names
dictionary + context
Not built
CVR and CPR numbers
deterministic checks
Not built
Case and file numbers
format per court and authority
Not built
Case content without clear markers
case-scoped routing + per-client blocking
Not built

The legal detector is not built. Confidentiality covers more than personal data, so the answer is two things: entities the engine finds, and case-scoped routing that keeps one case in one lane. The engine, the signed evidence layer and the personal-data core are validated in healthcare.

The rules it would be built for

Supervision: The Danish Bar and Law Society and the Danish Data Protection Agency

The Administration of Justice Act

a lawyer's duty of confidentiality covers the whole client relationship, not only personal data

The code of conduct for lawyers

confidentiality as a core duty

GDPR

client data, often special categories in family, criminal and immigration cases

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

Where lawyers would use it

  • Drafting letters and pleadings
  • Reviewing large document sets
  • Research with cloud models without client details in the request
  • Case-scoped AI assistants per client

Other sectors

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.