Enabling large language models for automated compliance checking of BIM models in architectural engineering through semantic modelling and linking

Purpose This study examines how standardised Building Information Modelling (BIM) data modelling can support large language models (LLMs) in automated compliance checking of architectural project information requirements. The focus is the Level of Information Need (LOIN) for design deliverables. The study addresses the disconnect between LLM reasoning and standardised project information. Design/methodology/approach A constructive research approach is adopted to develop and evaluate an LLM-based compliance checking framework. BIM resources and LOIN-based requirements are transformed into semantic data through information modelling and linking. An Information Container for Linked Document Delivery (ICDD) organises project documents and semantic links. The Model Context Protocol (MCP) connects LLM interaction with graph querying, compliance checking and issue reporting services. The prototype uses data from a single architectural project. A controlled benchmark compares direct review, Graph Retrieval-Augmented Generation (RAG) and MCP review. Findings Direct review based on raw Turtle snippets showed limited performance because compliance evidence was often interpreted conservatively. Graph RAG achieved higher checking performance through structured graph retrieval, while MCP review achieved the best overall result by combining LLM-based task interpretation with tool-mediated checking. Performance varied across rule types. Explicit-data and derived-value rules were checked more reliably than extended data-structure rules, which were sensitive to alternative relation patterns. Originality/value This study presents a prototype approach to BIM–LLM integration for compliance-oriented project information management. Project resources, information requirements and checking rules are organised through ICDD and semantic graph data. MCP-mediated interaction links requirement interpretation with executable checking services, evidence retrieval and issue reporting, enabling traceable compliance review in project delivery.

Authors

Institutions

Publication Details

Journal
Engineering Construction & Architectural Management
Published
2026-10-31
DOI
https://doi.org/10.1108/ecam-09-2025-1438
Primary Topic
BIM and Construction Integration
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Enabling large language models for automated compliance checking of BIM models in architectural engineering through semantic modelling and linking

Shihang Zhang, Siteng Zhang
Engineering Construction & Architectural Management
BIM and Construction Integration
article

Enabling large language models for automated compliance checking of BIM models in architectural engineering through semantic modelling and linking

Shihang Zhang, Siteng Zhang
article en

Abstract

Purpose This study examines how standardised Building Information Modelling (BIM) data modelling can support large language models (LLMs) in automated compliance checking of architectural project information requirements. The focus is the Level of Information Need (LOIN) for design deliverables. The study addresses the disconnect between LLM reasoning and standardised project information. Design/methodology/approach A constructive research approach is adopted to develop and evaluate an LLM-based compliance checking framework. BIM resources and LOIN-based requirements are transformed into semantic data through information modelling and linking. An Information Container for Linked Document Delivery (ICDD) organises project documents and semantic links. The Model Context Protocol (MCP) connects LLM interaction with graph querying, compliance checking and issue reporting services. The prototype uses data from a single architectural project. A controlled benchmark compares direct review, Graph Retrieval-Augmented Generation (RAG) and MCP review. Findings Direct review based on raw Turtle snippets showed limited performance because compliance evidence was often interpreted conservatively. Graph RAG achieved higher checking performance through structured graph retrieval, while MCP review achieved the best overall result by combining LLM-based task interpretation with tool-mediated checking. Performance varied across rule types. Explicit-data and derived-value rules were checked more reliably than extended data-structure rules, which were sensitive to alternative relation patterns. Originality/value This study presents a prototype approach to BIM–LLM integration for compliance-oriented project information management. Project resources, information requirements and checking rules are organised through ICDD and semantic graph data. MCP-mediated interaction links requirement interpretation with executable checking services, evidence retrieval and issue reporting, enabling traceable compliance review in project delivery.

Engineering Construction & Architectural Management
Tianjin University (CN), China Architecture Design & Research Group (China) (CN)
Openalex Percentile: Top 16%
BIM and Construction Integration
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon