THE QUESTION IT ANSWERS
What is this model allowed to change about itself?
- ACCOUNT
- PROTECT
- REPAIR
01
The problem
A deployed medical AI model is actively decaying in the field. Reapproval of a new model costs more than the original development. So it stays deployed, degrading, until it fails an audit. The FDA's Predetermined Change Control Plan (PCCP) framework creates an approved path — but only if the model's changes can be declared, bounded, and evidenced in advance.
02
What it delivers
Supervises model retraining strictly inside a declared, cryptographically signed bound. Produces a continuous, immutable record demonstrating that every internal modification stayed inside the declared regulatory limit. Built precisely to the FDA PCCP framework specifications.
03
Who it is for
A fit for
- Regulated clinical AI deployers facing model decay with no viable change-control path
Primary territory: Europe first (EU AI Act, Health Data Space mandate March 2029).
04
How it works
The Update Engine is built from 3 of the six capabilities of the ORMAS architecture. Each is a property of how the network trains, not a tool bolted on afterwards.
- 01ACCOUNTEmit a bounded, timestamped, causally attributed record of every modification the network made to itself.
- 02PROTECTGate a component out of the forward path when its health falls, so a failing region cannot poison what follows.
- 03REPAIRApply a bounded, conserved, logged correction to a named component under a declared per-diagnosis bound.
05
In your industry
This is why ORMAS is one product rather than seven. The Update Engine is the same mechanism in every row below — only the customer's own vocabulary changes.
- Medical
- supervised retraining
- Finance
- model recalibration
- Biotech
- assay-batch update
- AI lab
- scheduled refresh
06
Markets using it
- Regulated FinanceSegments by industry
- Medical AISegments by industry
- Defense & Safety-CriticalSegments by industry
07
Questions
No. ORMAS is one product under one licence, and The Update Engine is one of the seven things it is proven at. You turn on what your situation needs; there is no separate purchase, no separate integration and no separate contract.
The mechanism does not know what industry it is in. What changes between them is the noun, not the method — the table above is the same mechanism with each sector's own vocabulary substituted, which is why one product serves all of them.
On your infrastructure. The licence is on-premise and per-institution: nothing is sent to a third party, there is no hosted tier, and no usage metering phones home. In a hospital enclave or a controlled environment, that is a requirement rather than a preference.
Pre-book ORMAS with The Update Engine in scope.
Every engagement starts with one specific problem — a run that died, an audit that asked a question you could not answer, a dataset you cannot vouch for.
We reply within a few days. Every message is read by a person.
Pre-book a deployment
One licence, no price list. Terms are agreed per institution and written into the contract, and pre-booking fixes them ahead of general availability.
Useful to bring
- What you train, and roughly at what scale
- The obligation driving it — SR 26-2, the EU AI Act, Article 17, a PCCP, an internal policy
- What a failed run currently costs you to diagnose