Statute Engine
The statute book as living data, not static PDFs.
Working days to consolidate an amended Act
Cross-reference error rate, down from ~20%
Authoritative point-in-time statute book
The problem
The ministry held the country's entire body of statutes as static PDFs and Word files spread across departmental drives. Every amendment triggered a manual reconciliation: a drafter had to locate each downstream provision the change touched, hand-edit the consolidated version, and hope nothing was missed. Consolidating a single amended Act averaged 47 working days, roughly one in five published consolidations shipped with a cross-reference error, and citizens, courts, and firms were routinely citing superseded text because no authoritative live version existed.
In short
A legislative drafting and consolidation platform for a national ministry of law. We treated legislation as structured data rather than documents: every Act, section, and clause became an addressable, versioned node, and every cross-reference became a typed link. Amendments are modelled as operations applied to the graph, so consolidation becomes automatic, traceable, and citable instead of a manual reconciliation exercise that ran for weeks. The client is not named here in line with the confidentiality this engagement required.
- Consolidation of an amended Act cut from ~47 working days to under 3
- Cross-reference error rate reduced from roughly 20% to below 1%
- A single authoritative point-in-time statute book courts began citing directly
How it was built
- 01Model legislation as data.Every Act, section, sub-section, and clause became an addressable node with a version history, and every cross-reference became a typed link rather than a string of text. The statute book stopped being a pile of documents and became a queryable graph.
- 02Build the amendment engine.Amendments are modelled as operations (insert, substitute, repeal, renumber) applied to the graph. Feed in a passed bill and the engine maps its operative clauses to the target provisions and produces a consolidated draft with a full audit trail of what changed and why.
- 03Resolve every cross-reference.A resolver detects and flags broken or orphaned references the moment an amendment is applied, so the errors that used to surface months later in a courtroom are caught at drafting time.
- 04Render any date.Effective-date logic lets any provision be rendered as it stood on any date, so a court, a firm, or a citizen can cite the law as it applied to the facts in front of them.
- 05Publish and govern.A public point-in-time portal exposes citable permalinks, while role-based drafting workspaces give departmental drafters clause-level comments and approval gates so the authoritative version stays under proper control.
Under the hood
Results
- Consolidation time fell from an average of 47 working days to under 3, absorbing a 30% rise in legislative volume with no new headcount.
- The cross-reference error rate dropped from roughly 20% to below 1%, verified against a manual audit sample.
- For the first time the government published a single authoritative live version of its statutes, which courts began citing directly.
- Conservatively the department recovered on the order of 4,000 drafter-days per year, well over a million dollars in loaded staff cost, while removing a category of legal risk that had no price tag but real consequences.
Questions
