.png)
Predictive Maintenance in FM: Moving From Break-Fix to Data-Led Asset Management
- Maxcene Crowe
- Apr 8
- 6 min read
Reactive maintenance is not a system failure. It is a choice — and in most FM contracts, it is the wrong one.
When an asset breaks before it is touched, the costs run wider than the repair bill. There is the emergency callout, the tenant complaint, the compliance gap, the knock-on effect on other assets. All of it predictable. Almost none of it inevitable.
The FM industry has had the tools to move beyond break-fix maintenance for years. The real question is not whether predictive maintenance works — the evidence is clear. It is why so few contracts are structured to enable it from day one.
The 4 Maintenance Approaches: Where Does Your Contract Sit?
Before predictive maintenance can be implemented, it helps to understand where most FM operations actually sit on the maintenance spectrum.
1. Reactive (Break-Fix)
Action only taken after failure. High emergency costs, unpredictable downtime, poor audit trail. Still the default mode for a significant portion of FM contracts.
2. Preventative
Time-based servicing, typically aligned to manufacturer schedules or statutory compliance cycles. Better than reactive, but generates unnecessary maintenance activity when assets are performing well and misses failures between service windows.
3. Condition-Based Monitoring
Maintenance triggered by observed or measured performance thresholds. Requires sensor data or manual inspection regimes. A step towards predictive, but still retrospective in response.
4. Predictive Maintenance
Uses real-time sensor data, historical performance records, and AI or machine learning analysis to anticipate failures before they occur. Maintenance is scheduled when it is actually needed — not too early, not too late. ServiceChannel describes this precision as reducing overservicing, extending asset lifecycles, and lowering maintenance costs while cutting unplanned outages.
Most FM contracts operate somewhere between reactive and preventative. The goal of this post is to show what it takes to reach level four — and how to set that capability up from contract mobilisation.
What Predictive Maintenance Actually Requires
Predictive maintenance is not a product you can buy and install on day one. It is a capability built from four components working together.
Data. The foundation of any predictive model is historical asset performance data: failure history, maintenance records, energy consumption trends, work order patterns. Without baseline data, there is nothing to model against.
Sensors and IoT Integration. Real-time monitoring requires physical sensors on critical plant and equipment — HVAC, chillers, pumps, electrical distribution systems. These sensors feed continuous data streams that AI models can analyse for early fault signatures. As of 2025, 35% of facility managers report that IoT-driven predictive maintenance has reduced downtime by more than 20%.
Baseline Establishment. Predictive models need to know what "normal" looks like before they can detect deviation. This baseline period — typically 3 to 6 months — is where the predictive capability is being built, even if it is not yet being used.
CAFM Integration. Sensor alerts without connected workflows are noise. Effective predictive maintenance routes condition data into the CAFM or CMMS, triggering planned work orders automatically. IFMA confirms that AI systems which analyse data from building management systems and IoT devices enable facility managers to predict equipment failures and schedule maintenance during planned downtime — reducing disruption and improving asset lifespan.
The Mobilisation Implication
This is the point most contracts miss entirely.
Predictive maintenance capability cannot be retrospectively added to an FM contract at year two or three. The data requirements, sensor infrastructure, CAFM configuration, and asset baseline work all need to be scoped during mobilisation — ideally before handover from the incumbent.
The mobilisation phase is where the data architecture is set. Get it right, and predictive maintenance becomes an operational reality within 12 months. Get it wrong, and the contract defaults to the reactive maintenance model by operational inertia — regardless of what the bid document promised.
At transition, the priority actions are:
Obtain asset register and maintenance history from the outgoing contractor
Audit condition and sensor readiness for critical assets
Configure CAFM to capture condition data fields from day one
Define which assets are priority candidates for predictive monitoring
Mobilisation is not just logistics. It is the opportunity to lay the data infrastructure that the entire contract's maintenance strategy depends on.
What the Data Says
The business case for predictive maintenance is well established — and the numbers are compelling.
70–85% of a building's total lifecycle costs sit in day-to-day operations, making maintenance strategy one of the highest-leverage decisions in FM (Dexterra Group)
IoT-driven predictive maintenance has reduced downtime by more than 20% for 35% of facility managers (Dexterra Group)
A hospital using vibration analytics for early fault detection reduced diagnostic equipment maintenance costs by 25% (Dexterra Group)
A 10% reduction in energy costs can result in a 16% increase in profit margin — and predictive energy management directly drives that reduction (City FM)
As of 2025, only 28% of organisations have embedded AI solutions in their FM operations — meaning the competitive window for early adopters remains wide open (Dexterra Group)
Infraspeak notes that proactive identification and early fixes are demonstrably less expensive than reactive repair at a later stage — not as a theoretical claim, but as a measurable operational reality
The contrarian truth? Most FM contracts are structured to reward reactive response — callout charges, emergency uplift rates, materials markups. Predictive maintenance removes the financial incentive for the contractor to wait for failure. That tension needs to be designed out of the contract from day one.
A Realistic Implementation Roadmap
Predictive maintenance does not happen overnight. Here is a three-phase approach that works within the realities of FM contract timelines.
Phase 1: Foundation (Months 0–6, Mobilisation and Stabilisation)
Complete and verify asset register with condition ratings
Identify Tier 1 assets (critical to operations, high cost of failure)
Confirm CAFM can capture condition data; configure if not
Install or audit sensors on Tier 1 assets
Establish data collection baseline — failure history, energy consumption, maintenance frequency
Define KPIs: MTBF, MTTR, planned vs. reactive maintenance ratio
Phase 2: Analytics (Months 6–12)
Begin running condition monitoring data through CAFM analytics or third-party tooling
Build initial predictive models against Tier 1 assets using baseline data
Run parallel tracking: compare model-flagged maintenance against observed asset performance
Train operational team on alert interpretation and work order routing
Review and refine sensor thresholds
Phase 3: Operationalisation (Month 12+)
Transition Tier 1 assets to full predictive maintenance scheduling
Expand sensor and monitoring coverage to Tier 2 assets
Integrate predictive data into contract performance reporting
Use maintenance cost and downtime data to evidence contract value
Review annually and incorporate new asset classes
Saveable Framework: The Predictive Maintenance Readiness Checklist
Use this at mobilisation to assess whether a contract is structured for predictive capability.
Data Readiness
Complete asset register received from incumbent or built from site survey
Maintenance and failure history obtained (minimum 12 months preferred)
Asset criticality tiering completed (Tier 1 / Tier 2 / Tier 3)
Infrastructure Readiness
Sensor coverage confirmed or scoped for Tier 1 assets
IoT data feed connectivity confirmed or planned
BMS integration or data export capability confirmed
CAFM Readiness
CAFM configured to receive and log condition monitoring data
Work order auto-routing from condition alerts tested
Planned vs. reactive maintenance reporting enabled
Contract Readiness
KPIs defined: MTBF, MTTR, planned maintenance %, reactive maintenance %
Maintenance strategy documented and agreed with client
Escalation process for predictive alerts defined
Data ownership and reporting format agreed
Team Readiness
Operational team trained on condition monitoring tools
Escalation responsibilities assigned for predictive alerts
Reporting cycle for asset health agreed with client
Save this for your next mobilisation.
Academy Courses
If you are preparing a team to mobilise predictive maintenance capability from contract day one, these two courses are worth your time.
Covers the structured approach to contract mobilisation, including asset data, team setup, and the operational foundations that determine whether a contract performs from month one or spends a year recovering from a poor start.
Picks up where mobilisation ends — transitioning from setup mode to operational performance, embedding the processes and data disciplines that make predictive maintenance a live capability rather than a PowerPoint aspiration.
Comments