Use design, procurement and handover decisions to create maintainable assets and reliable data for future AI-enabled operations.

Professional illustration of a Singapore commercial facility with engineers reviewing a digital building model, tagged equipment, sensors and maintenance workflows.

For many building owners and facility managers, predictive maintenance starts too late. By the time an AI platform is being evaluated, equipment may already be difficult to access, asset tags may be inconsistent, sensors may not be installed and maintenance records may be spread across drawings, spreadsheets and emails.

Singapore’s upcoming CORENET X submission milestone creates a useful opportunity to move this conversation upstream. BCA has announced that, from 1 October 2026, new projects with a gross floor area of 5,000 square metres and above will be required to submit through CORENET X. At the same time, BCA’s Design for Maintainability findings reinforce a practical point: decisions made during planning and design affect future maintenance effort, manpower and lifecycle cost.

The objective is not simply to produce a compliant digital submission. It is to ensure that the building entering operations has maintainable equipment, usable asset information and a clear foundation for smart facilities management.

Why maintainability must be designed before handover

Maintenance teams inherit the consequences of design and procurement decisions. A pump may be technically suitable but difficult to isolate. An air-handling unit may be installed without sufficient service clearance. A warehouse system may depend on highly automated equipment without a clear maintenance access strategy. An electrical or mechanical asset may be documented in a way that does not match the labels used on site.

These issues create more than inconvenience. They can increase troubleshooting time, complicate permit-to-work planning, delay work orders and reduce the quality of data available for future analysis.

Design for Maintainability addresses these concerns earlier by asking how assets will be inspected, isolated, repaired, replaced and monitored throughout their operational life. The BCA study on Design for Maintainability links such decisions with potential lifecycle cost and manpower savings. For owners, the business case is therefore broader than construction quality: maintainability affects operational resilience and the total effort required to run the facility.

What CORENET X changes in the project conversation

CORENET X should not be treated only as a submission-system change. It is also a prompt for project teams to clarify information requirements, coordination responsibilities and digital deliverables before construction and handover.

For projects within the announced scope and timing, owners and developers should confirm how CORENET X requirements will be incorporated into the project information-management plan. Consultants, contractors and specialist vendors should understand which models, attributes and supporting information must be created, checked and transferred.

This is an opportunity to define operational information early, including:

  • Which maintainable assets must be represented in the digital model or asset register.
  • Which equipment attributes are required for operations, safety and maintenance planning.
  • How asset identifiers will match physical labels, drawings, schedules and work-order records.
  • Who is responsible for validating information before handover.
  • Which data formats and interfaces will be usable by the owner’s FM or computerised maintenance system.

The exact information requirements will depend on the project, systems and contractual arrangements. The important principle is to agree the operational use of information before asking teams to deliver it.

Five practical design decisions for AI-ready maintenance

1. Specify maintainability, not just equipment performance

Equipment schedules and specifications should address service access, isolation points, replacement routes, inspection requirements and maintainable component visibility where relevant. A design review should include the people who will operate and maintain the facility, not only those checking installation and performance.

2. Create a consistent asset identity

Every maintainable asset should have a clear identity that can connect the physical item to its model, location, manuals, warranty information, preventive-maintenance plan and future work orders. Agree naming conventions and tag structures early. Avoid allowing different contractors to create unrelated identifiers for the same asset.

3. Plan sensor access and data quality

AI diagnostics depend on usable time-series data. During design, identify where temperature, vibration, energy, pressure, run status or environmental information could support condition monitoring. Consider sensor mounting, power, network connectivity, calibration, cybersecurity, maintainability and data ownership.

Not every asset requires continuous monitoring. A practical approach is to prioritise equipment based on criticality, failure impact, operating conditions and the value of earlier intervention.

4. Design for automated work-order workflows

A detection is only useful if it can lead to an appropriate action. Asset data should support links between alerts, standard operating procedures, inspection checklists, responsible teams and escalation paths. This helps future systems automate routine work-order creation while keeping human review for decisions that require judgement.

5. Make handover a tested information process

Digital handover should not be a final file drop. Before practical completion, the project team should test whether an FM user can search for an asset, confirm its location, open the relevant document, understand its maintenance requirements and create or update a record.

Sample audits can check missing attributes, duplicate tags, incorrect locations, unreadable labels, outdated manuals and inconsistent equipment descriptions. The owner should accept information based on operational usability, not only file presence.

A pre-handover checklist for owners and FM teams

Before the building enters operations, owners, developers, consultants and FM representatives can ask:

  • Can the FM team identify every critical maintainable asset on site?
  • Do physical labels match the asset register, drawings and digital model?
  • Are access, isolation and replacement requirements documented?
  • Are equipment manuals, warranties, commissioning records and maintenance instructions linked to the correct asset?
  • Can the selected FM or maintenance platform receive the required data without extensive re-entry?
  • Are sensor points, gateways, network dependencies and data owners documented?
  • Have alarm priorities, response responsibilities and escalation rules been agreed?
  • Have safety controls and technology-enabled work processes been considered for maintenance activities?
  • Has the information been tested by actual operations or maintenance users?

For warehouses and logistics facilities, the review should also consider automated storage and retrieval systems, conveyors, charging infrastructure, environmental monitoring, vehicle movements and safe access around equipment. MOM’s WSH technology resources identify practical technology categories such as electronic permit-to-work, IoT environmental sensors, video analytics, robotics and vehicle-safety technology. These technologies should be introduced with clear operating responsibilities and safe work processes.

Where AI and automation can add value later

BCA’s AI for the Built Environment resources identify applications including AI-enhanced BIM, documentation, data management, workflow automation and facilities management. These applications are more effective when the underlying information is structured and trusted.

With suitable data, an FM team may be able to use automation to:

  • Prioritise inspections based on asset criticality and operating condition.
  • Detect unusual energy or equipment behaviour for human review.
  • Generate draft work orders from defined alarm conditions.
  • Retrieve manuals, procedures and past maintenance records through a controlled knowledge interface.
  • Identify repeated faults and support root-cause investigations.
  • Track response times, recurring failures and maintenance backlog trends.

These are not automatic outcomes of having a BIM model or installing sensors. They require data governance, integration, cybersecurity, suitable workflows and people who can validate AI-generated recommendations. For SMEs, a phased approach may be more practical: begin with a clean asset register and priority equipment, then add targeted automation and analytics.

What project teams should do now

Owners and developers should define the operational information requirements in the employer’s requirements and procurement documents. Consultants should include maintainability and asset-information reviews in design coordination. Contractors and vendors should confirm how equipment data, tags and manuals will be produced and validated. FM teams should participate before construction is complete, when changes are still possible.

A useful first step is a short readiness workshop covering critical assets, maintainability risks, CORENET X information responsibilities, handover requirements and future automation priorities. The output can become a project-specific data dictionary, asset-tagging plan, sensor strategy and handover test plan.

Singapore’s move towards more coordinated digital submissions and smarter operations makes this planning increasingly important. The best time to prepare AI-ready maintenance data is before poor information and difficult access become permanent operational problems.

ISS supports engineering, facility management and AI automation requirements for businesses that want to connect physical operations with usable digital workflows. Contact ISS to discuss your project, facility or maintenance-data requirements.

Reference points