Beyond barrier lists: GenAI-assisted systems and network-based diagnosis of Arctic off-site housing delivery
Purpose This study aims to develop a GenAI-assisted, source-traceable diagnostic to identify, prioritise and map interconnected constraints in Arctic off-site housing delivery within construction 4.0. Design/methodology/approach Following preferred reporting items for systematic reviews and meta-analyses 2020, 401 English-language peer-reviewed and grey-literature sources underwent a five-pass evidence extraction protocol, yielding 3,617 evidence records; a stratified sample of 360 was independently validated. The analysis produced a 36-code Arctic off-site construction challenge constraint framework, a multi-criteria challenge criticality index and a directed network of 36 nodes and 451 unique non-self-loop edges. Edge weights count unique supporting documents, while centrality describes unweighted network structure. Sensitivity testing assessed ranking stability, and family-level net-flow analysis examined directional patterns. Findings The indexed evidence base identifies volatile project costs, housing shortages and seasonal logistics as the most critical challenges. Network analysis reveals a different dimension of importance: seasonal logistics occupies the strongest bridging position, while shortages of senior northern project leadership rank third in betweenness centrality but only 17th in criticality, highlighting a potentially overlooked structural vulnerability in Arctic housing delivery. Research limitations/implications The network represents documented attributions rather than verified field causation and likely under-represents Inuktitut-language sources and community-held knowledge, highlighting the need for participatory validation. Practical implications The findings highlight candidate priorities for logistics-aware procurement, northern leadership development and community-led validation of governance–consent pathways to inform design for manufacture and assembly-oriented housing strategies. Originality/value This study integrates GenAI-assisted, source-traceable evidence extraction with criticality assessment and directed-network analysis. Unlike expert-elicited influence networks, relationships derive from documentary attributions, distinguishing documented prominence from structural position and revealing directional asymmetries.
Authors
- Ivanka Iordanova (ORCID: https://orcid.org/0000-0002-4596-2604)
- Amin Chaabane (ORCID: https://orcid.org/0000-0002-1113-7630)
- Sima Zarghami (ORCID: https://orcid.org/0009-0004-4490-6899)
Institutions
- École de Technologie Supérieure (CA)
Publication Details
- Journal
- Construction Innovation
- Published
- 2026-11-20
- DOI
- https://doi.org/10.1108/ci-03-2026-0105
- Primary Topic
- BIM and Construction Integration
- Type
- article
- Field-Weighted Citation Impact
- 0.00