Enhancing engineering quality: AI in practice

Engineers have always utilised software to improve productivity – from simple spreadsheets to finite element analysis software. Technologies continually evolve, enabling us to analyse structures faster, coordinate more effectively and deliver projects more efficiently than ever before. AI is the latest development in this evolution, with a key difference – its capabilities and flaws raise the question not just of what it can do, but of how it should be used.

At Heyne Tillett Steel, we believe AI has the potential to transform the way we work and we have been implementing AI tools in our day-to-day operations for some time now. In the built environment, where poor decisions can have serious real-world consequences, the use of AI must be approached with care. We firmly believe that the use of AI tools should enhance engineering quality rather than dilute it. However capable the technology becomes, human engineering judgement and accountability must remain at the centre of every decision.

We have always enjoyed developing our own software. The next generation of our in-house Carbon Counter will soon go live, and our free-to-use Stockmatcher tool has been used in more than ten countries to help engineers explore and implement reused steel on live projects. Our internal Site Visit Report App and Tech & Data reporting software are used daily, helping us to deliver consistently high-quality services more efficiently. It is in the same vein that we developed our own in-house AI portal, HTS-AI.

HTS-AI gives all HTS staff premium access to the leading AI LLMs (Large Language Models) and provides a secure space where we can develop and publish agents to automate specific tasks. Crucially, with this tool client data remains internal, outputs are controlled, and tools can be robustly tested before wider roll-out. Client data security and professional integrity are maintained.

Our approach to implementing AI focuses on delegating repetitive, admin-heavy tasks which do not require creative thinking or experienced judgement. Part of what we, and our clients, value so much is our ability to think critically and come up with interesting, creative engineering solutions. Holding onto this key skill within the company is central to our ethos as a practice.

We also use AI along with software API to help our multi-disciplinary engineering teams collaborate more effectively on complex projects, generating code to enable us to analyse models across different software packages, and scripts to retrieve and post-process the large amounts of data produced. Our structural and geotechnical teams, for example, co-produce soil structure interaction analyses to optimise the design of piled raft foundations and reduce pile loads, a task which wouldn’t be possible to do manually and without prior coding knowledge.

We are especially conscious of the impact on our younger and less experienced engineers. There is a risk that the increasing prevalence of AI tools can shortcut the formative experiences that develop good engineers while amplifying poor practice. Without careful guidance, the next generation may have fewer opportunities to develop the writing, communication and critical thinking skills that are essential in our line of work. We will always encourage younger engineers to develop their own understanding and judgement first, before utilising automation tools in an informed way. You can’t sense-check an output you don’t understand.

As a company, we are also conscious of the impact running these LLMs has on the use of energy, water, land and critical minerals. In parallel with our AI development, we are establishing processes for assessing our own ‘AI carbon footprint’, and investing time and resources into building ‘energy efficient’ agents, as well as providing training for our staff to ensure the models aren’t being misused.

So, how do our clients benefit from our development of these AI tools?

Time saved on repetitive and administrative tasks is reinvested into actual engineering. The goal is not efficiency for its own sake, but to create space for the deeper thinking that delivers better outcomes for our clients. Time saved can be better spent where it’s most needed: complex sequence planning, temporary works consideration, design innovation, spotting problems before they arise and getting stuck into the detail earlier to de-risk subsequent stages; these are just some of the ways in which the benefits of implementing AI tools extend to our projects.