
Building laws, connected in one AI assistant
A building-law assistant that brings national legislation, administrative rules, local ordinances and institutional replies into one place. It pairs explanations with source texts for practical questions about design, permitting and changes of use.
Regulations, ordinances & 1,911 replies
- Customer
- HAUD
- Areas
- Codes · Review
- Deployment
- Login-based web service · PoC

1Problem
Design and permitting reviews rarely depend on the Building Act alone. Questions about use, site conditions, parking and height require checking national legislation, subordinate rules, local ordinances and institutional replies together. Their sources and effective dates differ, so the material needs to be connected for review.
2Input
The collection brings together building-related national legislation, administrative rules, local ordinances and institutional replies. Users enter a design or permitting question and a reference date, adding a location or building conditions when needed.
3How it works
A PostgreSQL-backed collection keeps original texts, effective dates and editions distinct. Relevant regulations, ordinances and replies are supplied to the AI, while the deployed conversation interface separates explanations from source texts. Links at the reply date and the reference date remain separate, with uncertain dates and unresolved links left visible.
4Output
Users receive an AI explanation alongside related regulations and replies, with dates and source texts available for review. They can reopen or export saved conversations and leave feedback on issues they find. The final application still requires checking the design conditions against the original sources.
Results
- Preserved 687 institutional replies and added 1,224 from an official API, bringing the collection to 1,911. (As of October 21, 2026)
- Expanded to a PostgreSQL-backed collection covering building-law questions about design, permitting, use, parking and site conditions, with national rules and local ordinances within the collection scope.
- Made source provenance, dates and regulation-link status at different reference dates available for review.
- Deployed a web service with login, saved conversations, session resumption, original-source viewing and feedback collection.
Could this work on your projects?
Tell us about your challenges, and we'll explain how a similar approach could apply.
More case studies

AI aerial rendering for industrial plants
Scattered design data, restructured so AI can use it. Fine-tuned models and section-level prompts run as a lightweight deployment on the client's own infrastructure.
Phase 1 live, now in support and maintenance

Landscape options from site plans
Select an area on an early landscape plan and describe what you want. A model trained on Korean residential landscapes generates the options.
4 landscape options per area, in one pass


RFI answers for structural review
RFIs, structural drawings, and site photos, structured into one knowledge base. It finds similar past cases and drafts answers that cite the governing code clauses.
Draft RFI answers from similar past cases
