Singapore AfA-SHEL 2026: How AI and Digital Automation Can Make FM and Warehouse Work Safer

Practical AI, sensing and workflow ideas to prevent injuries, support returning workers and redesign demanding tasks.

Workplace safety is entering a more practical phase. Instead of treating digitalisation as a separate technology project, Singapore businesses can use AI and automation to improve how risks are identified, reported, controlled and reviewed.

This is especially relevant following the Ministry of Manpower’s open call for the Alliance for Action on Safety and Health for Employment Longevity, or AfA-SHEL. The open call closes on 31 August 2026. Selected organisations are expected to work on practical workplace-safety prototypes over approximately 14 months, with focus areas including injury prevention, returning to work after serious injury or illness, and workplace adaptation or job redesign.

For facility managers, warehouse operators, building owners and SMEs, the wider lesson is clear: safer work does not always require a large technology deployment. Well-designed pilots can reduce manual administration, improve response times and help supervisors make better decisions while keeping people responsible for final actions.

What AfA-SHEL means for FM and warehouse operations

Facility management and warehouse work often involve changing conditions. Workers may handle equipment, move loads, access plant rooms, work around vehicles, respond to faults or perform repetitive tasks under time pressure. The risk is not limited to a single unsafe act. It can arise from poor information flow, unclear handovers, unsuitable work design or a delay in escalating a near miss.

AfA-SHEL’s focus on employment longevity is therefore relevant to day-to-day operations. A safer workplace should help people remain productive over time, support workers returning after injury or illness, and adapt tasks where physical demands create avoidable risk.

AI can contribute by finding patterns and prioritising information. Automation can contribute by making the safer process easier to follow. Neither should replace competent supervision, worker consultation or proper risk assessment.

Five practical automation pilots to consider

1. Mobile-first hazard and near-miss reporting

A reporting process that requires lengthy forms may discourage timely reporting. A mobile workflow can allow a worker or supervisor to submit a photo, short voice note, location and basic category from a phone or tablet.

AI can help classify the report, identify repeated themes and route it to the appropriate person. For example, recurring obstruction reports in a loading area could be grouped for review rather than handled as isolated incidents. The system should support, not automatically determine, the severity of an event. A supervisor remains responsible for validating the report and deciding the response.

2. Automated escalation and close-out tracking

Safety actions can be missed when they sit across email, messaging applications, spreadsheets and paper records. A digital workflow can assign an owner, set a review date, record evidence and escalate overdue actions.

This is useful for common FM and warehouse activities such as housekeeping defects, damaged racking observations, blocked access routes, equipment faults or temporary controls. Dashboards should show open actions, ageing and recurring locations without turning the process into a punitive ranking exercise.

3. Sensor-based exposure and environmental monitoring

Environmental sensors can provide additional information about conditions that may affect work. Depending on the site, this could include temperature, humidity, noise, air quality, occupancy or equipment status. Wearable devices may also provide task or exposure information where there is a clear operational need and appropriate consent and governance.

The objective should be to identify conditions requiring attention, such as ventilation concerns, excessive heat or a congested work zone. It should not be continuous employee surveillance by default. Businesses should define what is collected, why it is needed, who can access it and how long it is retained.

For outdoor and physically demanding work, environmental information can support more timely decisions about breaks, task rotation, work sequencing or escalation. Any operational response should still follow applicable workplace-safety guidance and competent professional judgement.

4. Digital permits, method statements and shutdown workflows

Maintenance and engineering work often depends on accurate coordination. Digital permit-to-work and method-statement workflows can make required checks visible before work begins. They can prompt users to confirm isolation, access controls, affected stakeholders, emergency arrangements and handover requirements.

Automation can also help identify missing information or prevent a workflow from moving forward until a required step is completed. This does not make the system a substitute for authorised personnel. It reduces the chance that critical information is hidden in disconnected documents or missed during a rushed handover.

5. Safer job redesign and return-to-work support

AI and automation should also be applied to the design of work, not only to incident detection. A task review can combine observations, reported discomfort, manual handling requirements, equipment usage and workflow timing to identify where a job may be redesigned.

Possible interventions include introducing lifting aids, changing storage heights, reducing unnecessary walking, using remote inspection tools, splitting a task into stages or creating alternative duties during a return-to-work period. The right solution depends on the worker, task and workplace. AI may help organise information and suggest options, but decisions should involve the worker, supervisor and relevant safety or occupational-health professionals.

How to start a responsible pilot

1. Choose one defined risk: Start with a specific problem, such as delayed near-miss escalation, manual handling in a picking process or inconsistent maintenance handovers.
2. Map the current workflow: Document who reports, who reviews, what evidence is required and where delays occur.
3. Set a measurable operational objective: Examples include faster action assignment, higher near-miss reporting quality or improved completion of pre-work checks. Avoid claiming injury reduction before there is sufficient evidence.
4. Design human supervision into the process: Make clear who validates AI classifications, overrides recommendations and approves safety-critical actions.
5. Protect worker trust: Explain the purpose of data collection. Limit access, avoid unnecessary personal data and review privacy, security and retention requirements before deployment.
6. Test with frontline users: A technically capable system can still fail if it is slow, difficult to use or poorly matched to site conditions.
7. Review and scale carefully: Compare the pilot with the original process, capture unintended effects and expand only when the workflow is reliable.

Technology should strengthen safety culture

The best safety automation is not the most complex system. It is the system that helps workers report concerns early, gives supervisors useful context and makes agreed controls easier to complete.

Businesses should also consider basic operational resilience. Systems need clear ownership, reliable connectivity, backup procedures and a manual fallback for outages. An AI recommendation should never prevent a worker from stopping or escalating a task when conditions appear unsafe.

Singapore’s broader AI adoption direction, including support for SME capability building and pre-approved AI solutions, may make it easier for smaller organisations to explore practical use cases. However, technology selection should follow the operational problem, not the other way around.

A practical next step for Singapore businesses

Facility managers and warehouse operators can begin by selecting one process where safety information is currently slow, fragmented or difficult to act on. A small pilot in hazard reporting, environmental sensing, digital permits or job redesign can provide a more useful starting point than a broad, organisation-wide AI programme.

ISS can help businesses assess engineering and facility workflows, identify suitable automation opportunities and design practical digital services with human oversight. Contact ISS to discuss your engineering, facility management or AI automation requirements.