Aipedia: Architects

AI daylight, energy and overheating studies

Aipedia, by the Apex broking team · Last reviewed 2026-10-27

Architects can use AI-driven environmental analysis, but they must be able to explain and defend every assumption behind the result. Overheating and energy shortfalls are increasingly the subject of performance-gap claims against designers.

Part of: AI in Architects

In short

AI tools now estimate daylight, energy use and overheating risk from early models, supporting Part L (conservation of fuel and power) and Part O (overheating) work and daylight reports. The risk is a result nobody can explain: unvalidated tools, default weather or occupancy data, and inputs that drift from the final design. Under Part 2A reg 11K the designer must ensure the design would comply if built. When a building overheats or misses its energy targets, PI usually responds to negligent design advice, subject to wording, but contractual performance guarantees may not be covered.

How architects use AI for environmental analysis

AI-assisted tools can predict daylight levels, solar gain, energy demand and overheating risk from a massing or early design model, and suggest changes to glazing, shading or orientation. Practices use them for early design decisions, for daylight reports supporting planning applications, and as a first pass before formal Part L and Part O assessments.

Some tools are machine-learning surrogates trained on simulation results, giving fast answers that approximate a full simulation. Their accuracy depends on how closely the project resembles their training data, which is rarely disclosed. A result that looks reassuring at concept stage can therefore shape the whole design before anyone tests it properly, and early decisions on glazing and orientation are expensive to reverse.

The duties that apply

Building Regulations Part 2A, reg 11K: designers must ensure the design would comply if built, including with Parts L and O. Regs 11F to 11H set competence duties.
Architects Code Standard 3.1: work within competence. If the architect cannot explain the method, relying on it is risky.
Standard 3.2: keep up to date with current standards and guidance.
Standard 4.5: records of inputs, assumptions and the version of each tool.
Standard 5.4: tell clients promptly about issues affecting cost, time or quality, such as an assessment that changes once the full model is run.

The ARB and RIBA guide covers these standards in more detail.

What can go wrong

Illustrative scenario. At concept stage an architect uses an AI tool to show that a south-facing apartment block with large windows will not overheat. The client fixes the elevations on that basis. The formal overheating assessment later fails several flats, forcing smaller windows, external shading and a revised planning submission. The client claims redesign and delay costs, and argues that the architect's early advice was negligent.

After completion, a different claim may follow: residents complain of overheating or high energy bills and the client alleges the design did not perform as represented. See design claim examples. Building services engineers face parallel risks, set out in AI overheating assessments.

How PI usually responds

Negligent environmental design advice is usually within the scope of architects' PI, subject to the wording. Two points matter most:

Performance guarantees. PI commonly covers liability for failing to use reasonable skill and care, not guarantees that a building will achieve a particular result. An appointment promising specific energy or comfort outcomes may create liability the policy does not cover.
Who did the analysis. If the practice carried out work normally done by a specialist and this was not declared to insurers, cover may be questioned.

ARB's PII guidance sets the minimum at £250,000 each and every claim. See also does architects' PI cover AI mistakes.

Practical controls

Use AI results for direction only and confirm with a recognised assessment method before relying on them.
Check the weather, occupancy and ventilation assumptions against the project.
Rerun analysis when the design changes, and record which version informed each decision.
Avoid appointment wording that guarantees energy or comfort outcomes (see AI-drafted appointments).
Coordinate with the services engineer and record who is responsible for Part L and Part O.
Check compliance conclusions as described in AI building regulations checks.

Back to the AI in architecture overview.

How Apex helps

Apex arranges PI for architects. A broker can review how your policy treats performance-related claims and flag appointment terms that could fall outside cover.

Sources

Frequently asked

Can an AI daylight study support a planning application?

It can inform design, but planning daylight reports are usually expected to follow recognised methods. Check the council's requirements, use a validated method for the submitted report, and declare AI use where Planning Inspectorate rules apply on appeal.

Who is liable if a building overheats?

It depends on who designed the relevant elements and who carried out the assessment. Under Part 2A each designer must ensure their design would comply if built. Clear appointments and records of responsibility between architect and services engineer matter.

Does PI cover performance guarantees?

Usually not. PI commonly covers liability for negligence, meaning a failure to use reasonable skill and care. A contractual promise that a building will achieve a set energy or comfort result may go beyond that, depending on the policy wording.

How should AI tool results be recorded?

Record the tool and version, the date, the input model and the main assumptions such as weather data and occupancy. Note who reviewed the result and what decision it supported. This meets Standard 4.5 and helps defend later performance claims.

Check your PI covers how you use AI

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 callback

Apex 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.