Words: Annie Liang
Imagine a critical equipment failure in a major facility. The Computerised Maintenance Management System (CMMS) flagged the issue, but the timeline for resolution blew out due to a junior technician struggling with an unfamiliar and complex fault, as they lacked the nuanced diagnostic steps that an experienced engineer would
instinctively know. This costs time, money and reputation.
While the facilities management sector has embraced powerful technologies like CMMS and IoT sensors for efficiency and monitoring, these systems largely capture the outcome (what failed) rather than the invaluable process and insight of ‘how’ experienced human experts diagnose and solve complex issues.
A smarter AI strategy for facilities management doesn’t replace these experts with automation, but instead intelligently captures, structures and leverages their knowledge and real-time assessments to strengthen human capability, improve resilience and drive true predictive and preventative maintenance.
The current digital landscape: Data-rich, insight-poor
Australian FM operations widely use CMMS for work order management, IoT sensors for monitoring and various Software as a Service (SaaS) platforms for reporting and compliance.
These systems are effective at managing routine tasks, tracking Service Level Agreements (SLAs) and monitoring basic performance parameters. They handle structured data well, telling managers ‘when’ something happened and ‘where’.
The critical limitation is capturing the ‘dark data’ – the unstructured or semi-structured knowledge held by field experts.
Current systems can record that a pump failed (the ‘what’), but not the expert’s assessment of ‘why’ and ‘how’ it failed – the unique environmental factors observed, the subtle sound indicating the precise component fault or the creative workaround employed to minimise downtime (the ‘why’ and ‘how’).
This vital operational intelligence often remains fragmented – stored in voice notes, hurried sketches, mental models or brief and unstructured text fields.
The pitfalls of the wrong AI strategy
Many businesses fall into common AI adoption traps. Some of these include cost-cutting without addressing underlying process issues, creating new data silos with incompatible systems, and implementing AI that ignores or bypasses the human workforce.
Data silos prevent a unified view of operations and knowledge. Ignoring the human factor creates resistance, makes skill gaps insurmountable barriers and gives rise to ethical concerns, such as bias in data or perceived job displacement, that alienate the very staff who hold vital operational knowledge.
The untapped goldmine: Capturing expert assessments
A more human-centric AI strategy focuses on capturing and structuring the rich, often unstructured data generated by experienced FM personnel in the field.
Some examples include detailed voice notes describing observations during an inspection or repair; annotated photos of equipment, highlighting specific wear patterns or anomalies; structured forms capturing diagnostic steps and reasoning; video recordings of complex procedures with voiceover commentary; and text summaries explaining the context behind a
standard repair.
When processed by the right AI tools, this information delivers:
• A real-time, evolving knowledge base: A constantly updated repository of practical expertise, accessible organisation-wide.
• Enhanced predictive maintenance: Moving beyond sensor data (e.g., temperature spike) to incorporate expert judgment (‘that humming sound before the temperature spike indicates bearing stress’) allows for more accurate, context-aware
failure predictions.
• Powerful training tool: New technicians can access real world examples and expert reasoning for common and complex issues, addressing critical skill gaps and labour shortages in the industry.
• Instant knowledge retrieval: AI-powered search allows managers or field staff to query the knowledge base using natural language (‘What caused the last AC unit failure in Building 3, Level 5?’ or ‘How did John fix the recurring motor vibration issue?’) and retrieve relevant expert insights immediately.
• Improved compliance and accountability: Detailed capture of expert assessments provides robust documentation of actions taken, supporting compliance reporting and demonstrating due diligence under Australian regulations.
• Differentiation from standard CMMS: While CMMS handles workflow and structured data (the ‘what’ and ‘when’), these AI tools focus on the content and context of human observations and actions, building a layer of operational intelligence on top of existing systems. They provide the ‘why’ and ‘how’, not just the ‘what’ and ‘when’, safeguarding valuable assets and critical operations.
Looking ahead: Putting people at the centre of AI
Australia’s FM sector faces growing regulatory scrutiny, skills shortages and pressure to deliver more with less. In the next few years, the winners won’t be those who automate blindly but those who use AI to capture, enhance and share human expertise at scale.
For facility managers, the path forward is clear: invest now in AI that complements people – not replaces them – and transform field know-how into a resilient, intelligent asset. By doing so, FM leaders will reduce miscommunication and disputes, deliver faster resolutions and build a future-ready workforce prepared to meet evolving local compliance and operational demands.
This article first appeared in FM magazine.
