What is a Digital Twin of Legislation?
A DTL is a machine-executable, human-verifiable representation of a statute. It does not paraphrase the law — it mirrors it: every rule, rate and deadline stays linked to the provision that defines it.
From paragraph to executable decision.
Legal Text
The relevant provisions, quoted verbatim and anchored with European Legislation Identifiers (ELI).
Ontology
A formal semantic model (OWL) of every legal term, subject and parameter — queryable by humans and machines.
Configuration
Concrete rates, thresholds and dates as data, each annotated with its legal source — change the law, change the data, not the code.
Logic
Executable decision code with tests that reads every parameter live from the configuration and writes a cited derivation log.
Why “twin”, not “chatbot”?
A language model alone guesses. A twin computes: the logic is deterministic, the parameters are sourced, and every output explains which paragraph produced it. That is the difference between a plausible answer and a decision a public body can stand behind.
- Deterministic results — the same case always yields the same decision
- Full traceability — derivation logs cite ELI-anchored provisions
- Versioned — every published twin is pinned to an immutable snapshot of the law
- Updatable — when sources change, the twin detects it and flags affected layers
Standards under the hood
DTLs build on open standards: ELI for legal identification, LegalDocML classifications, OWL/Turtle for semantics, REST and the Model Context Protocol (MCP) for integration. No lock-in — the twin is an artifact you own.
Deterministic twins, built for agentic frameworks.
The purpose of a deterministic DTL is integration: agentic AI frameworks — Anthropic Claude, OpenAI ChatGPT, Google Gemini or Microsoft Copilot — connect to the twin via the Model Context Protocol (MCP). The assistant takes the human question about the law; the twin computes the answer — algorithmically, deterministically and certified against the legal text.
- Any MCP-capable assistant connects to a published twin in one line
- Citizens and case workers ask in natural language — the answer is computed by certified legal logic, never guessed
- Every answer carries a derivation log with ELI citations and the pinned version of the law
- getSemantic and explainDecision let agents query the legal ontology and justify any past decision
Curious how this looks for your domain?
We will walk you through a live twin — from legal text to API call — in 30 minutes.