Professional infographic showing a Singapore commercial building and warehouse connected through a central Smart FM workflow, with simple icons for assets, IoT sensors, BMS integration, digital work orders, cybersecurity and performance KPIs.

Smart Facilities Management is no longer just about installing sensors or adding another dashboard. For Singapore building owners, facility managers, warehouse operators and SMEs, the more important procurement question is how different systems, people and processes will work together in daily operations.

A Smart FM project can connect building management systems, IoT devices, asset registers, maintenance workflows, analytics and AI-assisted decision support. However, disconnected tools can create more administration, duplicate data and unclear accountability. A strong procurement specification should therefore define the operational problem first, then set clear requirements for integration, data governance, cybersecurity, service delivery and measurable results.

This is particularly relevant as organisations manage ageing assets, manpower constraints and growing expectations for productivity and sustainability. BCA’s recent Smart FM and facilities management resources emphasise the combination of systems, processes, technologies and people for data-driven operations. The following framework can help organisations evaluate solutions before committing to a platform or pilot.

1. Start with the operating model, not the technology

Before writing an RFP or requesting demonstrations, document how facilities work is currently performed. Map the process from issue detection to resolution and reporting.

  • How are faults reported: phone, messaging, email, paper or an existing system?
  • Who validates the issue and assigns the work?
  • Which team owns inspection, repair, approval and closure?
  • Where are asset records, warranties, manuals and maintenance histories stored?
  • Which tasks are delayed because information is incomplete or spread across systems?

This process map helps distinguish a genuine automation requirement from a feature request. For example, an organisation may not need a complex AI model at the start. It may first need a reliable asset register, mobile work orders and a clear escalation workflow.

2. Define the minimum digital foundation

Every Smart FM procurement should specify the information that must be captured consistently. At minimum, consider an asset register containing asset ID, location, equipment type, criticality, service history, responsible party and relevant documents.

Specify how assets will be identified across sites and systems. A chiller, pump or warehouse door should not have different names in the BMS, contractor records and work-order platform. Ask vendors how they will support data cleansing, duplicate removal, asset hierarchy and version control during implementation.

Also define the required records for digital work orders. These may include priority, fault category, assigned person, response time, completion time, parts used, photographs, notes, approval and root-cause classification. Consistent fields make later reporting and AI-assisted analysis more useful.

3. Be precise about BMS, IoT and system integration

Do not accept the general statement that a solution can integrate with existing systems. Ask for an integration schedule that identifies each source system, data type, direction of data flow, update frequency, interface method, responsible party and testing approach.

Potential sources may include a BMS, energy meters, access systems, environmental sensors, equipment controllers, warehouse systems, contractor portals and finance or procurement platforms. The specification should state which data must be read, which actions may be written back, and which systems remain the system of record.

For IoT devices, clarify installation responsibilities, connectivity, power requirements, calibration, battery replacement, device health monitoring and handling of missing or abnormal readings. A sensor that stops reporting should generate an exception, not silently disappear from the dashboard.

Where a digital twin or visual building model is proposed, define its purpose. It may support asset visualisation, logistics, space management or operational decision-making, but the project should not pay for a complex visual layer without a clear workflow or business use case behind it.

4. Specify AI by decision and outcome

AI should be procured as part of an operational decision process rather than as a standalone feature. Examples include prioritising work orders, identifying unusual equipment behaviour, classifying service requests, summarising inspection notes or recommending next actions.

For each proposed AI use case, ask:

  • What input data is required and who is responsible for its quality?
  • What decision or action will the output support?
  • Who reviews, approves or overrides the recommendation?
  • How will false alerts, missed alerts and uncertain outputs be handled?
  • Can the organisation see the reason, evidence or data behind an alert?
  • How will performance be reviewed after deployment?

This approach avoids buying an AI label without a usable operating process. It also ensures that automation supports FM personnel instead of creating unreviewed decisions or additional alert fatigue.

5. Make digital workflow requirements concrete

A digital work-order system should reflect the organisation’s actual service model. Include requirements for mobile access, role-based views, approval steps, contractor assignment, recurring planned work, emergency jobs, escalation rules, attachments and audit history.

For field teams, consider whether the workflow works in plant rooms, loading bays, basements and other operational areas. The solution may need photo capture, simple forms, QR or barcode asset lookup, offline handling or clear status updates. These requirements should be tested in a real site environment rather than only in a conference-room demonstration.

For managers, specify the exceptions that need attention: overdue work, repeated failures, abnormal readings, critical asset downtime, unresolved complaints or missing compliance records. Exception-based management is more useful than a dashboard filled with information but no prioritisation.

6. Set data ownership and cybersecurity expectations

Data ownership should be written into the procurement documents. Clarify who owns asset data, sensor data, work-order records, user-generated notes, analytics outputs and configured workflows. Define access rights, export formats, retention periods and the process for returning data when a contract ends.

Request a clear explanation of the hosting model, user authentication, role-based access, audit logs, backup arrangements, vulnerability handling, incident notification and subcontractor access. Requirements should align with the organisation’s internal policies and any applicable contractual or regulatory obligations. Vendors should explain how data is separated between customers and whether customer data is used to train shared AI models.

Cybersecurity is not only an IT issue. Connecting operational technology and building systems can affect facility operations. Establish who approves network changes, who monitors device connectivity, who maintains credentials and who responds if an integration fails.

7. Allocate vendor responsibilities clearly

Many Smart FM projects underperform because responsibilities are assumed rather than documented. Your procurement pack should identify the owner for site surveys, data cleansing, device installation, network coordination, integrations, user training, testing, go-live support and ongoing service management.

Ask vendors to provide an implementation plan with dependencies and acceptance criteria. A useful acceptance test may check whether assets are correctly mapped, work orders reach the right team, alarms create the intended workflow, reports reconcile with source data and users can export required records.

Also distinguish between platform support and operational FM responsibilities. A technology vendor may maintain the application and integrations, while the FM team or service provider remains responsible for inspection, diagnosis, repair and escalation. The contract should make these boundaries visible.

8. Use KPIs that measure operations, not platform activity

Agree on baseline measurements before implementation where possible. Relevant KPIs may include work-order response time, completion time, overdue work, planned versus reactive maintenance, repeat faults, first-time resolution, asset data completeness, alarm-to-action time and user adoption.

For energy or sustainability use cases, define the measurement method, data source, baseline period and factors that could affect results. Avoid treating the number of sensors, logins or dashboard views as proof of business value. The most useful KPI is linked to a better operational decision or a measurable reduction in wasted effort, delay or risk.

9. Plan a phased implementation

A practical rollout often begins with one site, asset group or workflow. A first phase could establish the asset register, integrate a limited number of data sources and digitise priority work orders. The next phase can introduce exception rules, analytics, additional IoT coverage or AI-assisted recommendations once data quality and user adoption are understood.

Each phase should have a defined scope, owner, budget, timeline, training plan and go/no-go criteria. Review what worked before expanding to more buildings or contractors. This reduces the risk of scaling inconsistent data and ineffective workflows across the portfolio.

Procurement checklist

  • Document current FM processes and pain points.
  • Define the required asset, work-order and sensor data.
  • List every system to be integrated and identify the system of record.
  • Specify AI use cases, human review and exception handling.
  • Set data ownership, export and access requirements.
  • Request practical cybersecurity and operational technology controls.
  • Assign responsibilities for implementation, testing and support.
  • Agree on operational KPIs and baseline measurements.
  • Start with a controlled pilot and clear acceptance criteria.

Smart FM procurement is ultimately a business process design exercise supported by technology. By specifying how data moves, how work is assigned and how decisions are reviewed, Singapore organisations can avoid disconnected tools and build a more dependable digital operating model.

ISS can help organisations assess engineering, facility management and AI automation requirements, including workflow design, system integration and phased digital implementation. Contact ISS to discuss your requirements.

Further reading