FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: Building services engineers
AI can help build and run Part L compliance models, but the designer must still ensure the design would comply if built and the model matches what is installed. Compliance models that drift from the built design can cause delay at completion.
Part of: AI in building services engineering
In short
MEP and energy engineers use AI to populate SAP and SBEM-style compliance models from drawings, suggest specifications that pass, and write compliance reports. The risk is a model that passes on paper but no longer reflects the building, because inputs were assumed or the design changed. Part 2A regulation 11J requires designers to plan, manage and monitor their work, and regulation 11K requires the design to comply if built. Failed completion sign-off can lead to redesign, remedial work and delay claims, which PI usually treats as design negligence claims subject to the wording.
Part L of the Building Regulations covers the conservation of fuel and power, with guidance in Approved Document L for dwellings and for other buildings. Compliance is commonly shown using the national calculation methods: SAP for dwellings and SBEM for other buildings, or approved software built on them. AI tools now extract fabric areas and services data from models, propose combinations of fabric, heating and renewables that pass, run sensitivity checks and draft the compliance report.
Used well, this speeds up option studies. Used badly, it produces a passing model built on assumptions nobody has checked, which then becomes the benchmark the contractor and building control body work to.
On higher-risk buildings, regulation 38 of SI 2023/909 requires the client to hand over information including fixed building services and fuel and power use on completion, so modelling errors can carry into the golden thread.
Typical problems are AI-assumed thermal bridging or airtightness values, services efficiencies taken from generic data rather than the products specified, renewables counted that were later value-engineered out, and models never updated after design changes. Each can produce a compliance figure that cannot be reproduced at completion.
Illustrative scenario: an energy engineer uses an AI tool to find a passing specification for a small block of flats, accepting its assumed airtightness and heat pump efficiency. The contractor later substitutes plant and the as-built inputs no longer pass. Completion sign-off is delayed, the developer pays for additional measures and holding costs, and claims against the designer for the cost of the gap, arguing the original compliance strategy was never realistic.
The compliance model is only as good as its link to what is installed. On many projects the energy engineer hands over to a contractor who selects plant, and value engineering removes or changes items the model relied on. AI can help here too, for example by comparing submitted product data against modelled inputs, but someone has to own that comparison. Agree in the appointment whether you review substitutions, and if you do, treat every change to fabric, airtightness targets, heating plant, hot water or renewables as a trigger to re-run the model. If you are not asked to review substitutions, say so in writing, so the gap in responsibility is visible before completion rather than after.
Claims for redesign, remedial measures and delay arising from a flawed compliance strategy are usually treated as design negligence claims, and PI commonly responds subject to the policy terms. Questions can arise over who was responsible for updating the model after changes, and over contractual promises of particular ratings. Our comparison of civil, structural and M&E engineer PI explains how insurers view MEP risk alongside other disciplines, and our notes for energy assessors are useful if you also do assessment work.
Related: AI overheating assessments, AI net zero advice and AI and the golden thread. Back to AI in building services engineering.
Apex arranges PI for building services engineers. A broker reviews your cover and how your policy treats AI-assisted compliance modelling and any rating commitments in your appointments.
It can help populate and run the model, but the designer must ensure the design would comply if built under regulation 11K. Every assumed input should be checked and the model kept aligned with the design.
It depends on the appointments and on who controlled the changes. Designers have duties under regulations 11J and 11K, and contractors and clients have their own Part 2A duties, so responsibility is often disputed.
PI usually responds to claims alleging negligent design, which can include resulting delay costs, subject to the wording. Contractual promises of specific ratings may go beyond what a policy covers.
Yes. Keeping each version with a note of the changes helps show you planned, managed and monitored the design, and is useful evidence if a claim is made years later.
Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.
Start a PI quote Request a callbackApex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.
