What Technology Does to the Case for SFG20: IoT, Digital Twins, and the Future of Hard FM Scheduling
The economics of condition monitoring have changed faster than most Hard FM contracts have. A vibration sensor that cost several hundred pounds in 2015 now costs under £30. That single fact is quietly making SFG20 frequency scheduling redundant on the assets that generate the most task volume.
The Fundamental Assumption SFG20 Makes That Technology Is Challenging
SFG20's maintenance schedules are built on a single operating assumption: that we do not know the condition of an asset between service visits. Because we cannot observe it continuously, we service it at intervals calculated to prevent failure before it occurs. The interval is set conservatively — frequently enough that failure is unlikely, infrequently enough that the schedule is commercially viable. The interval is fixed regardless of whether the asset is new or old, lightly loaded or running at capacity, in a clean office or a dusty plant room.
This assumption was entirely reasonable in 1990, when SFG20 was created. Building management systems existed but were expensive, proprietary, and limited in scope. IoT sensors did not exist in any commercial sense. The only practical source of asset condition data was the engineer standing in front of the asset with a clipboard. Fixed-interval maintenance was not a limitation of the standard. It was the only rational model given the available information.
In 2026, that assumption is no longer universally true. A modern commercial office estate, an NHS trust with a current BMS, or a data centre will have continuous real-time data on the performance of its primary M&E plant: supply and return temperatures, flow rates, pressure differentials, energy consumption, vibration signatures on rotating plant, runtime hours, fault codes. The engineer with a clipboard is no longer the only source of condition data. On a well-instrumented estate, the BMS knows more about the assets between service visits than the SFG20 schedule assumes.
SFG20 was built for a world where fixed-interval maintenance was the only rational model. On a modern instrumented estate, it is no longer the only model. It is increasingly the expensive default.
Four Technologies That Are Directly Challenging the SFG20 Frequency Model
The cost of IoT sensor technology has fallen by roughly 80 percent over the past decade. A vibration sensor that would have cost several hundred pounds per unit in 2015 now costs under thirty pounds. A temperature and humidity sensor costs under ten. A wireless data gateway covering a plant room costs under two hundred. For a medium-sized estate, comprehensive IoT instrumentation of primary M&E plant is now a capital investment measurable in tens of thousands of pounds rather than hundreds of thousands.
What continuous IoT monitoring delivers that SFG20 cannot: anomaly detection before failure. A bearing on a pump motor that is developing a fault will show a characteristic vibration signature change weeks or months before it fails. A chiller running inefficiently due to refrigerant loss will show a degrading coefficient of performance in its energy consumption data before any engineer identifies it in an annual service visit. An AHU with a blocked filter will show increasing pressure differential across the filter bank. All of these are detectable continuously from sensor data. None of them are detectable from a fixed-interval service schedule until the service visit arrives.
The commercial consequence is specific: on assets where IoT monitoring is deployed, the SFG20 service frequency becomes a ceiling rather than a floor. The engineer visits when the data says to visit, not when the calendar says to visit. On assets that are performing well, visit frequency reduces. On assets showing early deterioration, frequency increases before the asset fails. Both outcomes reduce cost and improve reliability compared to the fixed-interval model.
Modern BMS platforms from vendors including Siemens Desigo, Honeywell Enterprise Buildings Integrator, Johnson Controls Metasys, Schneider Electric EcoStruxure, and Trend Controls do far more than monitor setpoints. Current generation BMS platforms generate fault codes, performance deviation alerts, energy consumption anomalies, and runtime-based maintenance triggers. On a well-configured BMS, the system can generate a maintenance work order directly when a piece of plant reaches a defined performance threshold, without any human intervention in the trigger process.
The gap between this capability and current Hard FM practice is significant. In most Hard FM contracts, BMS data sits in the client's or contractor's BMS system and is reviewed periodically by a BMS engineer or facilities manager. It is not systematically integrated into the CAFM PPM scheduling system. SFG20 schedules run in the CAFM. BMS alerts are managed separately. The two datasets rarely talk to each other in real time. The result is that estates are running both a fixed-interval SFG20 schedule and a BMS alert system in parallel, with manual processes connecting them. Neither is informed by the other in the way the technology now makes possible.
- BMS generates fault code. BMS engineer or FM manager reviews alert manually.
- Reactive work order raised manually. CAFM updated manually.
- SFG20 PPM schedule continues unchanged. No feedback loop.
- Annual service task executed regardless of BMS performance data.
- Outcome: parallel systems, manual bridges, missed signals.
- BMS monitors asset performance continuously against defined baselines.
- Deviation from baseline triggers a CAFM work order automatically.
- Work order includes BMS performance data for the attending engineer.
- Engineer's findings update the asset record and the BMS baseline.
- SFG20 schedule frequency adjusted dynamically based on performance history.
- Outcome: condition-triggered maintenance, evidence-based scheduling, reduced unnecessary visits, faster response to genuine deterioration.
The technology for the integrated model exists now. The contracting model, CAFM configuration, and procurement specification that would require it does not yet exist on most UK Hard FM estates.
A digital twin is a virtual representation of a physical asset or building, updated in real time from sensor and operational data, that can be used to simulate asset behaviour, predict failures, and optimise maintenance scheduling. Digital twin technology for the built environment has moved from a research concept to operational deployment over the past five years. Vendors including Siemens, IBM Maximo, Autodesk Tandem, and specialist built environment platforms are offering digital twin capabilities specifically designed for facilities and asset management.
The relationship between digital twins and SFG20 is not one of replacement but of obsolescence pressure. A digital twin of an AHU system continuously tracks the actual operating parameters of each component against its design specification. When a component's real-world behaviour diverges from its digital model by a defined threshold, the twin generates a maintenance recommendation. That recommendation is specific to the actual condition of the specific asset instance, not to the generic asset class that SFG20 addresses.
For an FM operator running a digital twin on primary plant, the SFG20 schedule for that plant becomes redundant as a scheduling tool. The schedule is replaced by the twin's recommendations. SFG20 may still inform what tasks should be performed during a visit, but it no longer determines when. The frequency model — which is the commercial heart of SFG20's value proposition — is removed from the equation.
This is not yet mainstream in UK Hard FM. Digital twin deployment on operational buildings is concentrated in new high-specification commercial developments, data centres, healthcare facilities with significant capital investment in BMS infrastructure, and government buildings under the Government Soft Landings framework. But the direction of travel is clear, and the question for Hard FM operators is how quickly the installed base of modern estates will reach the point where SFG20 frequency scheduling is the legacy model rather than the standard one.
The most significant near-term technology pressure on SFG20 is not IoT sensors or digital twins, which require upfront capital investment and instrumentation. It is AI-driven analysis of existing operational data: energy consumption records, reactive maintenance histories, BMS fault logs, and equipment runtime data that most estates already hold but rarely analyse systematically.
AI maintenance scheduling tools — including those being developed by CAFM platform vendors and specialist startups — ingest historical operational data and identify patterns that predict failure before it occurs. A chiller that has had three reactive callouts in eighteen months for the same fault code is telling the CAFM system something that the SFG20 schedule cannot detect: that the annual service frequency is insufficient for this specific asset on this specific estate. An AI system analysing that data history would flag the asset for a change in maintenance regime. The current SFG20 schedule would continue generating annual service tasks until the next failure.
The commercial case for AI-driven scheduling optimisation is straightforward: reduce unnecessary planned maintenance visits on assets with good performance histories; increase attention on assets with deteriorating performance patterns; identify assets where the SFG20 frequency is wrong for the specific operating context. Early adopters in the UK Hard FM market are reporting planned maintenance cost reductions of 15 to 25 percent on instrumented asset classes where AI scheduling has been deployed.
What This Means for Facilities-iQ as a Platform
Facilities-iQ's commercial position is built on a specific value proposition: it is the authoritative source of SFG20 task codes and frequencies, delivered through a subscription platform that FM contractors and clients pay to access. The commercial logic of that position depends on SFG20 task frequencies remaining the primary scheduling input for planned maintenance on UK estates.
Technology is eroding that dependency from the top of the asset criticality curve. On the highest-value, most-monitored assets, condition-based and AI-driven scheduling is replacing fixed-interval SFG20 frequencies. The assets most likely to be instrumented first are also the assets that generate the most SFG20 task volume: chillers, AHUs, pumps, BMS-integrated plant. As instrumentation spreads down the asset criticality curve, the proportion of the estate where SFG20 frequencies are the primary scheduling input will shrink.
| Asset Category | Current SFG20 Dependency | Technology Trajectory | SFG20 Relevance in 5 Years |
|---|---|---|---|
| Primary chillers, large AHUs | High: SFG20 schedule drives PPM | IoT and BMS integration already deployed on new and refurbished plant | Low: condition monitoring replaces frequency scheduling on modern plant |
| Pumps, fans, compressors | High: SFG20 frequency for all | Vibration and current monitoring viable at low cost per unit | Medium: legacy plant stays on SFG20; new plant moves to CBM |
| Electrical distribution, switchgear | High: SFG20 plus statutory | Thermal imaging and power quality monitoring increasingly standard | Medium: statutory tasks remain; discretionary frequency under pressure |
| Fire, life safety, emergency lighting | High: SFG20 plus statutory mandate | Technology assists evidence capture but cannot replace statutory visits | High: statutory obligation does not change with technology |
| Small plant: FCUs, radiators, VAVs | High: large task volume | Low-cost IoT viable but instrumentation of every FCU not yet commercial | High: SFG20 remains default for non-instrumented small plant |
| Fabric: doors, windows, decoration | Low: SFG20 covers M&E only | No technology pressure on fabric maintenance scheduling | High: SFG20 does not cover fabric; no change in that position |
The picture that emerges is not one where technology eliminates SFG20 across the board. It is one where technology eliminates SFG20 on the asset classes where it generates the most task volume and, by extension, the most subscription justification. The assets that remain firmly in the SFG20 model are the statutory and life safety tasks, the small plant that is commercially impractical to instrument, and the fabric maintenance that SFG20 does not cover in the first place.
For Facilities-iQ, this is a strategic challenge. The subscription model is justified by the value of SFG20 frequencies as scheduling inputs. As those frequencies become redundant on instrumented primary plant, the justification for the subscription narrows to statutory tasks and non-instrumented small plant. That is a smaller and less compelling value proposition than the current one. The platform will need to evolve to provide value in a condition-based world, or it will face the same obsolescence pressure it is currently applying to manual maintenance scheduling.
Facilities-iQ's commercial position depends on SFG20 frequencies remaining the primary scheduling input on UK estates. Technology is making that assumption less true every year. The question for the platform is whether it evolves before the erosion becomes structural.
What Hard FM Operators Should Do Now
The technology trajectory is clear. The timeline is not. The pace at which IoT, BMS integration, and AI scheduling will displace SFG20 frequencies across the general Hard FM estate depends on capital investment cycles, procurement specification evolution, and the rate at which CAFM platforms develop genuine real-time BMS integration rather than periodic data imports.
For Hard FM operators making decisions today, the practical position is this: technology does not yet replace SFG20 on most of the estate. But on specific asset classes and specific estate types, the investment case for condition-based monitoring over SFG20 fixed-interval scheduling is already compelling, and the operators who are building that capability now will be better positioned than those who are not when the market moves.
- Map your estate by instrumentation status. Which assets already have BMS integration or IoT monitoring? Those assets are candidates for condition-based scheduling today. The SFG20 schedule for those assets should be reviewed against the available performance data.
- Identify your CAFM platform's BMS integration capability. Can your CAFM receive real-time BMS triggers and generate work orders automatically? If not, what would it take? This is a contract and configuration question, not a technology question.
- Build your reactive maintenance history into your scheduling model. Which assets have the highest reactive callout frequency? Those assets are telling you the SFG20 schedule is wrong for them. AI tools can identify the pattern. A good contracts manager can identify it manually.
- Specify technology integration in your next retender. If your next Hard FM contract does not specify real-time BMS integration, AI-assisted scheduling review, and condition monitoring on primary plant, you are procuring a 2015 model in 2026. The market can deliver more.
- Do not wait for a technology transformation to fix the SFG20 basics. IoT and AI do not fix a wrong asset register, an inaccurate labour model, or a misunderstood compliance position. Technology built on a bad foundation produces bad results faster. Fix the foundation first.
The technology questions raised in this article intersect directly with Baachu's advisory practice. We help Hard FM operators understand:
- Which assets on their estate are ready for condition-based scheduling and what the business case looks like in their specific context
- What CAFM and BMS integration capability their current contract delivers and what should be specified in the next one
- How to use existing reactive maintenance data to identify SFG20 frequency failures before investing in new technology
- How to write technology requirements into Hard FM tender specifications that are measurable and commercially realistic
Baachu Rain tracks 11,000+ UK Hard FM contracts. We know what the market is specifying now and what the leading operators are doing differently.
Contact us: hello@baachu.com · baachurain.com
Frequently Asked Questions
Not entirely, and not on all asset classes. Technology is replacing SFG20 frequency scheduling on the highest-value, most-monitored assets — primary chillers, AHUs, pumps, and BMS-integrated plant — where condition-based monitoring makes the fixed-interval model redundant. But statutory tasks remain frequency-based by regulation regardless of technology, small non-instrumented plant is not commercially viable to monitor individually, and fabric maintenance sits outside SFG20 entirely. The realistic outcome for most UK Hard FM estates over the next five years is a hybrid model: condition-based monitoring on primary instrumented plant and SFG20 retained as the default for the rest. Technology does not eliminate SFG20; it shrinks the portion of the estate where SFG20 is the primary scheduling input.
A digital twin is a virtual representation of a physical asset or building, updated in real time from sensor and operational data, used to simulate asset behaviour, predict failures, and optimise maintenance scheduling. For Hard FM, a digital twin of primary plant continuously tracks actual operating parameters against design specifications and generates maintenance recommendations when real-world behaviour diverges from the model. The key impact on SFG20 is that the twin removes the frequency scheduling function: it tells you when to visit based on actual condition, not on a calendar. SFG20 may still inform what tasks to perform during a visit, but the when — which is SFG20's core commercial value — is replaced by the twin's output. Digital twin deployment is currently concentrated in high-specification commercial developments, data centres, and NHS facilities with significant BMS investment. It is not yet mainstream on the general commercial or public sector estate.
AI maintenance scheduling tools ingest historical operational data — energy consumption, reactive maintenance histories, BMS fault logs, equipment runtime data — and identify patterns that predict failure before it occurs. The key advantage over SFG20 is that the analysis is asset-specific and estate-specific: an AI system will identify that a particular chiller has a recurring fault pattern suggesting the annual SFG20 service frequency is insufficient, while also identifying that a group of FCUs with clean performance histories are being over-serviced relative to their actual deterioration rate. The result is a maintenance schedule shaped by actual performance data rather than generic frequency tables. Early adopters in UK Hard FM are reporting planned maintenance cost reductions of 15 to 25 percent on instrumented asset classes. The important caveat is that AI scheduling requires good underlying data — a clean asset register, complete maintenance histories, and integrated CAFM and BMS records. AI built on poor data produces poor results faster.
True BMS-CAFM integration means that when the BMS detects a performance deviation — a chiller operating outside its efficiency envelope, a pump showing abnormal pressure readings, an AHU generating a fault code — it automatically triggers a work order in the CAFM system, complete with the underlying performance data for the attending engineer. The engineer's findings then feed back into the asset record and update the BMS baseline. The SFG20 schedule frequency for that asset is adjusted based on its actual performance history. This is technically achievable with current platforms from major BMS and CAFM vendors. The reason it is not standard is contractual and commercial, not technical. Most Hard FM contracts do not specify real-time BMS integration. CAFM systems are configured to run SFG20 schedules, not to receive condition-based triggers. BMS and CAFM are managed by different teams, often different contractors. Procurement specifications written in 2018 do not require a capability that did not exist commercially in 2018. The gap will close as retenders specify integration requirements, but it will not close automatically on existing contracts.
No, and this is an important distinction. Statutory maintenance obligations — Legionella L8 monitoring and treatment, fire alarm testing and maintenance, emergency lighting testing, lift thorough examination, LOLER, PSSR, fixed wire testing, F-gas compliance — are mandated by legislation at specific frequencies. Technology can assist with evidence capture: digital records, automated logging, sensor-verified completion. But it cannot substitute for the statutory visit itself. A sensor that confirms water temperature is not a substitute for a Legionella risk assessment and treatment programme carried out by a competent person. An IoT-monitored fire detection system still requires the periodic testing and maintenance required by BS 5839. The technological transformation described in this article affects discretionary PPM on performance-monitored assets. Statutory obligations are a fixed floor that technology makes easier to evidence but does not make smaller.
- Art. 1 Is SFG20 Outdated? The Hard FM Baseline That Built an Industry and Why It Is Now Costing You Money
- Art. 2 SFG20 Labour Hours: Why Your PPM Pricing Is Wrong Before the Contract Starts
- Art. 3 Who Owns SFG20? BESA, Facilities-iQ, and the Hard FM Commercial Risk
- Art. 4 What SFG20 Compliance Actually Costs: Five Hard FM and TFM Perspectives
- Art. 5 The Asset Register Problem: Why the Foundation of Every Hard FM Contract Is Built on Data Nobody Has Verified
- Art. 6 Beyond SFG20: The Credible Alternatives for Hard FM Maintenance Strategy
- Art. 7 What Technology Does to the Case for SFG20: IoT, BMS Integration, Digital Twins and AI Maintenance Analytics (this article)
- Art. 8 How to Build a Defensible Hard FM Maintenance Framework Without SFG20 as the Anchor.
- Art. 9 SFG20 State of FM Report 2026: What the Data Actually Shows About SFG20 Compliance, Asset Registers and Hard FM Dependency.
Ready to map which technology applies to which assets on your estate? One email starts the conversation. → hello@baachu.com
Next: Article 8 · How to Build a Defensible Hard FM Maintenance Framework Without SFG20 as the Anchor
Read Article 8 →