________________________________________________________________________________________________________________________
the as-built drawings. The fixings might be non-compliant and cavity barriers missing. A material certificate has been issued for a product that was never installed. No algorithm ever catches that. It requires a site visit, careful observation and a willingness to ask searching questions.
The same investigative approach applies across residential and senior living work. Ground conditions shift in ways a desktop study won’ t capture. Previous alterations to a structure might not appear on the drawings. The construction schedule that works on paper may not work for the contractor on site. Weighing up those factors and applying appropriate remedies across a live project remains the responsibility of engineers.
The Building Safety Act reflects this reality, placing formal responsibility on engineers to exercise professional judgement and document the basis of every decision. That accountability continues long after the project completes and it cannot be delegated.
So where does AI belong? It handles much of the iteration work well. It helps with non-technical report drafting, navigating large volumes of guidance and checking documents. This frees up time for engineering professionals to make high impact judgement calls.
What AI can never do is replace the engineer who walks the site, reads the room in a client meeting and decides when output needs to be questioned rather than accepted. That questioning culture does not emerge by accident. It depends on who is in the room.
What women bring to engineering intelligence
According to research from EngineeringUK, only around one in six( 17 per cent) of UK engineers are women. That under-
The judgement gap
The industry talks often about a skills gap. There is also a judgement gap and it widens every time output from an AI tool goes unchallenged. This is where mentorship relates directly to AI.
Senior engineers need to involve junior colleagues in design discussions early, helping them understand the reasoning behind decisions. A junior engineer who knows why a decision was made is better placed to recognise when an AI model’ s output doesn’ t align with engineering judgement. Treating that knowledge transfer as core to technical delivery, rather than as a separate concern, is how teams stay sharp.
16