What is machine-executable regulation?
Machine-executable regulation is regulation whose requirements are expressed as formal rules that software can evaluate deterministically: given facts about a situation, the rules determine what is required, permitted or prohibited, and why. It goes beyond machine-readable regulation, which structures the text, by adding executable semantics, defined inputs and outputs, and provenance from each rule to the provision it implements.
What execution requires
Turning a requirement into something a system can run takes more than translation.
- A formal rule representation: obligations, permissions, prohibitions, conditions and exceptions in a language with precise semantics, such as LegalRuleML or a domain-specific language.
- Defined inputs: the facts, data fields or evidence the rule evaluates.
- Deterministic evaluation: the same inputs always produce the same result, which can be tested.
- Precedence and defeasibility: which rule wins when rules conflict, and when exceptions apply.
- Temporal validity: which version of the rule applied at a given time.
- Provenance: a link from the executable rule, and from each evaluation, to the source provision.
Deterministic validation around probabilistic AI
Large language models are good at interpreting text and poor at guaranteeing the same answer twice. Machine-executable regulation places formal rules around that interpretation: AI proposes the structure and candidate rules, people review them, and the rules then execute deterministically. The probabilistic step is contained; the decision step is reproducible and auditable.
Where machine-executable regulation is used
- Automated regulatory compliance: evaluating policies, controls and data against requirements.
- Decision support: generating reasoning paths with traceability to the relevant provision.
- AI governance: giving agents and pipelines explicit regulatory constraints they can check against.
- Public administration: implementing legislation in digital services consistently and transparently.
- Regulatory change management: re-evaluating affected rules and systems when legislation changes.
How NormLogic makes regulation executable
NormLogic turns structured regulatory knowledge into formal rules, validates them for consistency and completeness, and executes them with an audit trail: source, interpretation, rule, reasoning and decision. The rules remain linked to their provisions, so when the regulation changes the affected rules are identifiable.
Frequently asked questions
Is machine-executable regulation the same as a rule engine?
A rule engine executes rules; it does not know where they came from. Machine-executable regulation adds the connection to the source regulation, the structured knowledge in between, and the provenance that makes results auditable. NormLogic can feed rule engines but is not one.
Can all regulation be made executable?
No. Some requirements depend on judgement, proportionality or facts that are not available as data. Regulatory Engineering makes executable what can be executed, makes the rest explicit as structured knowledge, and records where human judgement enters.
How is machine-executable regulation tested?
Like software: with cases. Known situations with expected outcomes are run against the rules, and regressions are detected when rules or regulations change. Provenance makes failing cases traceable to the provision and interpretation involved.
Related topics
- What is machine-readable regulation?
Machine-readable regulation is regulation in a structured format software can parse: provisions, concepts and requirements with identifiers. What it includes.
- What is automated regulatory compliance?
Automated regulatory compliance evaluates policies, controls and data against machine-executable rules, with a traceable finding for every gap.
- What is regulation as code?
Regulation as code expresses legislation, regulation and policy in executable form. Its relation to rules as code and policy as code, and NormLogic’s approach.
- NormLogic, the platform
How Sireto applies Regulatory Engineering to machine-readable and executable regulation.
See a regulation execute.
Bring a requirement you currently check by hand. We will show the formal rule, the evaluation and the evidence trail.