BIMinAM AI and Data Engineering Workstream Meeting 05-02-2026

BIMINAM AI AND DATA ENGINEERING WORKSTREAM MEETING, 05-02-2026

Date: 5 Feb 2026
Focus: Using AI to improve digital handover, asset data quality, and Golden Thread delivery

Purpose of the Session

  • Explore practical use of AI to improve digital handover, asset information, and Golden Thread outcomes.
  • Move from exploratory discussion to usable workflows, tools, and collaborative outputs.
  • Focus on as-built information quality, validation, and lifecycle data continuity across planning → creation → operation.

Key Problems Identified

  • Low trust in handover data and widespread gaps in Regulation 38 / O&M information.
  • Asset data often delivered as disconnected:
    • Models
    • COBie
    • PDFs and O&M manuals
    • 360 photos
  • Poor version control, classification, and taxonomy alignment.
  • Clients’ information requirements often unclear or unenforced.
  • AI currently used to “fix” unstructured data rather than working from structured inputs.

Golden Thread & Data Principles

  • Golden Thread described as a continuous, machine-readable information DNA.
  • Spaces and assets remain the core structuring mechanism.
  • Standardisation is essential to enable automation and interoperability.
  • AI should support:
    • Validation
    • Structuring
    • Auditing
    • Integration
      not replace proper information management.

Practical AI Use Cases Demonstrated

  • Automated auditing of large handover document sets (e.g. 10,000 files).
  • Extracting schedules and asset data from:
    • PDFs
    • Drawings
    • Door schedules
    • Apartment layouts
  • Identifying missing or inconsistent data across models and O&M.
  • Rapid production of validation reports to return to supply chain.
  • Use of prompts and pipelines to turn manual QA processes into automated checks.

Key Risks & Governance Issues

  • Liability and PI risk when using AI to validate third-party information.
  • Need for clear provenance and traceability of:
    • Source data
    • Prompts
    • Outputs
  • Risk of using design data instead of as-built data.
  • Need for human oversight and ISO 19650-aligned check/review/approve processes.

Client & Commercial Drivers

  • Clients must better understand:
    • Legal duties
    • Data ownership
    • Information requirements
  • Clear AIR/EIR enforcement seen as critical.
  • Group exploring commercial model for collaborative delivery teams using shared methods and tools.
  • Benefits framework being developed to demonstrate cost and risk savings.

Agreed Direction

  • Form a focused AI/automation working group (“automate”).
  • Share prompts, schemas, and validation methods across members.
  • Avoid duplication of effort across organisations.
  • Build repeatable, standardised workflows for Golden Thread delivery.

Next Steps

  • Collect examples of AI tools and use cases from participants.
  • Develop shared prompt libraries and validation frameworks.
  • Run follow-up sessions (extended to 90 minutes).
  • Continue linking AI work to Golden Thread benefits and business case.
  • Engage more clients and supply-chain partners.

Overall Conclusion

  • AI offers significant potential to improve handover quality, compliance, and operational insight.
  • However, value depends on:
    • Standardised data
    • Clear responsibilities
    • Client leadership
    • Collaborative implementation
  • The workstream will now focus on practical, testable outputs rather than theory.