Safety Advisory and Guidance Engine

The Safety Advisory and Guidance Engine (SAGE) is an AI-powered workplace-safety prototype announced by the Workplace Safety and Health Council at the Singapore Workplace Safety and Health Conference on 23 July 2026. The WSH Council developed it with the Singapore Institution of Safety Officers and industry partners, including partners engaged through Built Environment Action Teams. SAGE is intended to help workplace-safety professionals access relevant information more quickly while complementing—not replacing—professional judgement (MOM, 23 July 2026).

Announced prototype functions

The prototype was described as having three capability areas: analysing workplace photographs to identify common safety risks; recommending relevant guidance based on similar scenarios; and providing language support for migrant workers. The intended benefit is faster access to relevant information for workplace-safety professionals, including across language contexts, rather than a new category of statutory inspection. These are announced prototype functions, not evidence that SAGE has achieved a particular accuracy level or that every workplace can already use it (MOM, 23 July 2026).

Testing and expected availability

The WSH Council and SISO planned to invite workplace-safety professionals to test the prototype over the following months before a full launch expected in early 2027. The expected timing is a forward-looking implementation target, not proof of a completed public deployment or an open self-service tool for employers. Current users should rely on the WSH Council's live guidance and professional assessment until SAGE's operating status, access conditions and safeguards are published (MOM, 23 July 2026).

Relationship to workplace-safety duties

SAGE is an information and decision-support tool, not an automated compliance decision-maker, statutory safety standard, or substitute for risk assessment, supervision and employer duties under Singapore's workplace-safety regime. Its output should be reviewed against the facts of the worksite and the applicable legal and technical requirements. It is also distinct from MAST, which coordinates multi-agency safety responses, and AfA-SHEL, which co-creates practical solutions for injury prevention, return-to-work and safer job redesign (MOM, 23 July 2026).

Intended users and evidence boundary

The announced user group is workplace-safety professionals, whom the WSH Council and SISO planned to invite to test the prototype before its expected early-2027 launch (MOM, 23 July 2026). The speech presents SAGE as a way to surface relevant information more quickly: a photograph may help identify a common risk, similar scenarios may point users toward guidance, and language support may improve access for migrant workers. It does not publish an accuracy rate, the training-data composition, a list of supported languages, a public-access route, or a date on which employers can rely on it for compliance decisions. Those omissions are material: a prototype recommendation is a prompt for human review, not proof that a hazard has been comprehensively detected or that a workplace has satisfied its legal duties.

The policy context is therefore capability-building within the existing WSH system. MOM's speech places SAGE alongside leadership, worker reporting, training, safe-work procedures and supervision, and says its role complements rather than replaces professional judgement (MOM, 23 July 2026). Until the operating model and safeguards are published, questions about liability, auditability, data retention, model performance and formal enforcement should be answered from the applicable WSH legislation and live Council guidance, not inferred from the prototype announcement.