Automated Design Compliance Checks for Internal Guidelines

Functional Requirements

• Read Revit (.rvt) and/or ArchiCAD models submitted by consultants
• Check design models against CLD’s internal guidelines and highlight non-compliances
• Map internal rules (from SharePoint) into machine-readable checks, and automatically verify compliance against these rules (including geometry, naming conventions, annotation, and layout constraints)
• Allow CLD to update and version-control guidelines centrally
• Identify non-compliant items and generate structured reports showing: what failed, where it failed (location in model), and which guideline it violates
• Generate feedback in near real-time (within hours)
• Handle both 2D and 3D drawings
  

Technical Requirements

Core Requirements:

• Must accept Revit (.rvt) and/or ArchiCAD models directly (not IFC)
• Allow guidelines (from SharePoint or other structured formats) to be updated and converted into machine-readable rules
• Handle rule types such as:
  • Parametric checks (e.g., corridor width ≥ 1.2m)
  • Spatial relationships (e.g., window-to-window distance)
  • Element classification (e.g., naming standards for fire stairs)
  • Level-based zoning (e.g., designated plant area not exceeded)
• Generate interactive feedback and summary reports with annotated markups, checklists, and references to guideline clauses
• Provide either a Revit plug-in (for in-model checking) or an external web-based submission portal (for consultant uploads)
• Must be robust enough to process large BIM models and handle multiple asset classes

Additional Preferences:

• A rule-based or hybrid AI engine that converts design guidelines into a structured rule library
• A BIM model parser that can query and verify geometry, metadata, and naming structures
• A feedback engine that overlays results directly on model views or exports annotated reports
• A central admin dashboard to update, version-control, and manage rulesets across projects
• Generative AI capability to suggest improved layout or design optimisations within the parameters of CLD’s design guidelines, replacing the current process where design managers manually test alternatives before recommending revisions
• Multi-language support (especially Chinese) for overseas projects
• Ability to evolve into a compliance dashboard with historical learning

‍Expected Outcomes

• Reduction of time required to check consultants’ designs for compliance with CLD’s internal guidelines – from the current 1 to 2 full working days (8.5 hours per day) per proposal to at least 30 minutes, representing a time saving of over 90% for each round of checking
• Reduction of manual compliance checking effort by at least 50% across the 3 to 5 design iterations typical of a project, accelerating design review cycles
• Further time savings if Generative AI can also propose improvements and optimisations
• Automatic detection of 80-90% of compliance issues, reducing risk of human oversight and error
• Early detection of non-compliance to prevent downstream redesign, delays, and redundant work
• Improved quality through consistent, traceable QA for internal guidelines
• Enable CLD’s design managers to focus on higher-value activities, such as design research and exploring new ideas and innovations that can add value to developments

Deployment Environment and Constraints

The solution will be applied to large and complex design submissions in Revit or ArchiCAD. Consultants’ familiarity with plug-ins or external portals may vary. The system must be robust enough to handle large BIM models and scale across multiple asset classes and regions. Turnaround time should be quick enough (just hours) to align with design iteration cycles.

CLD will support a pilot using live project BIM models. Support includes:

• Sample project models with known deviations from guidelines
• Access to current guideline documentation in SharePoint
• Feedback from internal design reviewers
• Validation against live project timelines for turnaround and accuracy

‍Commercialisation and Scaling

Pending a successful pilot, CLD intends to implement the solution at its various country offices across internal design workflows, as a required pre-check for all consultants’ early design stages. This rollout will cover multiple project types and could be extended to other business units involved in design review. CLD intends to scale the solution across projects regionally in Singapore, Vietnam, and China.