Published on :
June 11, 2026
by
Anisha Bhattacharjee
For most FM organisations, the first clear answer comes within 90 days.
Not from a deployment. Not from a system change. From a structured diagnostic on data they already own. Work order history, maintenance records, asset performance data sitting inside existing CMMS, CAFM, and BMS systems. Analysed through a governance lens, that data can surface whether autonomous maintenance is viable, where the gaps are, and what the operational and financial case actually looks like.
What happens after those 90 days depends entirely on what the evidence shows.
Autonomous Maintenance is the operating model in which AI governs routine maintenance decisions across PPM scheduling, reactive work order management, SLA compliance, and asset reliability, while human FM teams retain responsibility for governance and exception management.
This is not automation of tasks. It is governance of decisions.
In an autonomous environment, AI handles anomaly detection, work order classification and prioritisation, PPM interval recalibration, reactive maintenance routing, SLA monitoring, and maintenance outcome verification. Human supervisors no longer coordinate every individual maintenance activity. They manage exceptions and govern the system that operates on their behalf.
Most FM organisations today are managing maintenance reactively. Faults are logged after they occur, PPM schedules are calendar-based rather than condition-driven, and supervisors spend the majority of their time coordinating execution rather than governing outcomes.
The shift to autonomous operations does not happen through a single deployment. It happens through a structured transition, starting with understanding what your existing data already reveals, then progressively moving decision-making from human execution to AI governance.
The starting point is not a technology commitment. It is an evidence question: does autonomous maintenance actually work on our assets, with our contractors, and within our operating model?
That question is answered before any live system integration begins.
Xempla is an AI-native operational governance platform purpose-built for facilities management. It sits above existing CMMS, CAFM, IWMS, and BMS environments as a decision layer, not replacing the systems FM teams already use, but governing the decisions those systems currently leave to manual coordination.
Your systems remain the systems of record. Xempla becomes the system of decisions above them, turning operational activity into governed, accountable outcomes.
The entry point into that journey is the 90-Day Accelerator.
The 90-Day Accelerator is a paid diagnostic. Not a pilot. Not a proof-of-concept.
It is a structured engagement that answers a more fundamental question: what is your FM operation actually delivering today, and where are decisions being made without sufficient evidence?
Most FM operations run on activity records. Work orders closed, inspections completed, SLAs met. What they cannot answer is whether any of it is working. The Accelerator surfaces the operational truth your existing data already contains.
It takes a historical data export from the systems you already operate, your CMMS, CAFM, BMS, or ERP, and runs it through Xempla's governance engine. No live system access or API integration is required. It does not replace or modify your existing FM systems, require new sensors or hardware, change contracts or SLAs, or commit either party to a broader rollout.
At the end of 90 days, you receive eight named deliverables and a genuine choice about what comes next.
The Accelerator runs across five domains. Maintenance Effectiveness, Reactive Load Reality, Risk and Reliability Exposure, Energy and Control Behaviour, and Decision Traceability. Examining decision quality and outcome assurance in each.
Day 1 — Secure Data Intake Data received via secure upload from existing CMMS, CAFM, IWMS, BMS, or ERP systems. No live connection established. No disruption to operations.
Days 1 to 30 — Governance Mapping Evidence extraction and Operational Truth baseline built from existing data.
Day 30 — First Briefing Operational Truth Dashboard, risk scores, and maintenance effectiveness baseline.
Days 30 to 90 — Full Assessment Spend versus outcome alignment mapped. 3 to 5 year scenario model built.
Day 90 — Full Governance Briefing Eight named deliverables. Board-ready output. Three options for next steps: maintain current approach, address specific gaps, or explore broader operating model change. No obligation either way.
For the full breakdown of deliverables and what the Accelerator examines, see the 90-Day Accelerator page.
At the core of Xempla's operational governance model is the DIIV Cycle, the framework through which every maintenance event is analysed, acted on, and verified, whether running retrospectively on historical data during the Accelerator or continuously on live inputs post-integration.
Discover AI identifies deviation from expected asset behaviour using BMS inputs, sensor data, and historical fault patterns.
Investigate The deviation is examined to understand root cause, failure pattern, and intervention priority.
Implement An intervention is generated and executed or routed for human approval, depending on action type and defined risk threshold.
Verify Outcome is confirmed against the SLA baseline and MTBF target. The result feeds back into the asset's Assurance Score and recalibrates future intervention timing.
If the Day 90 assessment justifies proceeding, live API connections are established to existing CMMS, CAFM, IWMS, and BMS environments. The governance model transitions from historical snapshot to continuous live stream. The DIIV Cycle begins running in real time, detecting anomalies, generating intervention recommendations, and verifying SLA closure as it happens. Same platform. No migration. No re-onboarding.
What this looks like in practice:
One of the UK's largest FM service providers across Health and Care PFIs started with the 90-Day Accelerator on a complex, predominantly reactive portfolio. The proof justified live deployment. Today, one reliability engineer governs eight PFI sites, with a 30 to 40% reduction in reactive callouts and 100+ critical downtimes avoided. Read the full case study here.
For most FM organisations, the first clear answer comes within 90 days, through a structured diagnostic on existing CMMS, CAFM, and BMS data. No live integration is required. At Day 90, the organisation has a definitive, evidence-based picture of whether autonomous maintenance is viable on their own assets, and a genuine choice about whether to proceed.
The 90-Day Accelerator is Xempla's paid diagnostic programme that analyses historical FM data to surface the operational truth an organisation's existing reporting cannot produce, across five domains, with eight named deliverables at Day 90. No live system integration is required to begin. If the results justify proceeding to live deployment, that decision is made at Day 90 with full evidence. If not, the engagement stands alone as a completed operational assessment.
No. The 90-Day Accelerator runs entirely on a historical data export from existing CMMS, CAFM, IWMS, BMS, or ERP systems provided via secure upload. No API connections, workflow changes, or operational disruption are required to begin.
Eight named deliverables: an Operational Truth Dashboard, Maintenance Effectiveness Assessment, Reactive Load Analysis, Decision Readiness Summary, Energy and Control Insight Brief, Risk and Lifecycle Indicators, a 3 to 5 Year Scenario Model, and a Recommendation on Next Steps with three clearly presented options and no obligation to proceed.
The DIIV Cycle, Detect, Implement, Intervene, Verify, is Xempla's framework for AI-governed maintenance decisions. It runs retrospectively on historical data during the 90-Day Accelerator and continuously on live data post-integration, governing how AI identifies asset deviation, generates interventions, routes them for approval, and verifies outcomes against SLA and MTBF targets.
At maturity, AI governs fault detection, work order classification, PPM scheduling, contractor routing, escalation management, and SLA verification within defined controls. Human teams manage governance, exception handling, and strategic recalibration. The operational model shifts from coordinating maintenance activity to governing the system that coordinates it.
Paragraph
Block quote
Ordered list
Unordered list
Bold text
Emphasis
Superscript
Subscript
© Xempla. All rights reserved.
Terms & Conditions