The challenge of AI adoption in the built environment sector

In early October a group of 40 leaders and senior executives from across the built environment sector gathered in London to discuss the impact of artificial intelligence (AI) on the sector’s business model and what leaders should do next to maximise its potential.

The roundtable followed on from an executive survey undertaken globally by Pinsent Masons, Bentley Systems, Mott MacDonald and Turner & Townsend, both to better understand how the built environment sector is currently adopting AI and examine the likely investment in emerging use cases in the built environment. We also wanted to understand the readiness and maturity of the sector in adopting AI as well as the future risks, challenges and opportunities that the sector faces in doing so.

The group included leaders from across government, as well as investors, legal experts, technology solutions providers, project and cost managers, engineering designers, contractors, clients and developers engaged in the built environment sector.

The consensus industry view is that AI is largely only expected to have an incremental impact on the built environment sector to start with, but that significant disruption should be anticipated when the benefits of AI become more widely felt across the sector.

The adoption of AI across the sector will require significant levels of investment to change a siloed and heavily fragmented approach to a more integrated, modern and digital, data-driven business model.

The effective adoption of AI across the sector will also require significant structural change in the way in which the sector is currently organised and operates. The sector is traditionally conservative and slow to adopt technology. Other sectors, such as aerospace and automotive, are also highly regulated and safety-critical, but have already experienced structural and business model change that has enabled the fast take-up of technology, and with it, more rapid and deeper improvements in productivity and competitiveness. More integrated supply-chains and an industrialised approach to construction will help create step-changes in productivity and efficiency.

We are currently exploring how organisations in the built environment sector across different countries are adopting AI and what stages they are at in terms of maturity of adoption.

The global tech sector may yet prove to be a disrupter. It is investing trillions of dollars into AI infrastructure globally, supporting a global race that will reshape the world and every sector of the global economy.

How AI will impact the industry's business model

AI has the potential to significantly disrupt the built environment sector. Building regulations, codes and standards around the world demand that those designing and constructing infrastructure and buildings take responsibility for safety and integrity as well as other factors, such as accessibility to people with disabilities; energy efficiency; and, increasingly, carbon footprint and resources used.

A total of 40% of all businesses across the built environment sector in our survey anticipate that AI-driven change will impact their current business model to a significant extent. A further 24% believe AI-driven change is a major consideration for their businesses and are taking active steps to adapt their business model in anticipation of significant future disruption.

Sector readiness and maturity

There’s growing evidence that the built environment sector still lacks the necessary readiness and maturity to successfully adopt AI in the short-term.

It’s clear, both from our survey and our own industry experience, that the sector is nascent in its readiness to develop, adopt and use bespoke AI solutions.

The highly fragmented nature of the construction industry and the localised nature of construction delivery are also key barriers to the sector adopting technology more broadly.

Investment in AI has been sunk into automating document-related processes and optimising construction through improved planning. There is a much greater current focus on optimisation of design and engineering and improving cost estimation and forecasting. Continued investment in cost estimation and forecasting and real-time tracking of site progress looks to be a particularly strong focus.

We believe that the automation of site activities is coming down the track, but it’s clear that digitalisation of engineering design and the standardisation of data throughout the supply-chain must happen first.

A significant shift in mindset and approach is required to increase the pace of adoption within the industry. Currently, the use of bespoke AI tools and agentic AI within the industry is limited. It is acknowledged that there is a huge knowledge gap in the industry, with particular unease regarding the use of agentic AI systems – which leverage ‘static’ AI agents and other tools and resources to build additional capabilities, like reasoning, planning and self-evaluation – in the sector. Concerns about the potential impact the technology may have on jobs is also compounding the reluctance to explore what the technology can achieve.

While adoption of AI solutions is still at a nascent stage within the built environment sector, it is widely acknowledged that AI is an “unavoidable force”. However, wider uptake of the technology will only be achieved once barriers to adoption are overcome. Barriers to take-up include:

• limited investment ‘dry powder’: typical profit margins in the construction industry leave little headroom for investment compared to other sectors. Construction companies therefore want to ensure that a solution works before investing in it, otherwise it could potentially be a very costly mistake;
• fragmented and localised nature of the industry: take-up and adoption of AI is dependent on a range of factors including the location, size and readiness of the business. Differing standards and regulations from region to region and sector to sector add to the complexities of standardising AI models;
• supply chain and continuity of work: construction projects are not a continual production line, but there are stops and starts between projects. This is a contributory factor in a fragmented supply chain and a lack of continuity. In addition, designs, specialisms and products vary from project to project, which also adds to the increasing complexity of supply chains;
• concerns around data security, model bias and trust are a significant barrier for a risk-averse industry. In addition, operational siloes, a lack of standardisation and machine-readable data can all hinder adoption
• AI literacy and leadership are very important factors in the adoption of AI solutions. For AI projects to succeed, the industry needs to invest in people, workforce and skills to create an AI-ready organisation.

As AI becomes an increasingly integral part of the construction industry, it will have a significant impact on procurement and business models. An executive explained that their organisation has been working on an AI solution where a task, which used to take three months, now only takes three hours when using AI. How the industry costs work will need to be reevaluated as AI becomes more embedded in construction processes.

Better integration of supply chains is often talked about but, in reality, very little has changed in practice. However, AI could be the key to driving increased virtual integration in the construction industry by consolidating processes, improving data-driven decision-making and enabling organisations to have greater control over the whole value chain.

Risks for organisations in the built environment sector

Industry transformation through growth in AI will give rise to opportunities, but could also lead to increased risks for organisations.

Governance and risk management

The adoption of AI is likened to “gold rush fever”. By having a clear vision of priorities and having an AI adoption roadmap in place, businesses can remain focused on achieving tangible outputs. The risk of making expensive mistakes is heightened without a robust governance framework in place.

Our survey highlighted a lack of robust governance regarding the use of AI. Although over half of organisations had adopted organisational AI policies, only 20% had gone further and implemented organisational AI policies that set out guidelines for use, governance, ethical implications, safety measures and related aspects.

Health and safety

Health and safety is a critical issue in the construction industry, and the use of AI could increase the risk of mistakes and liability. For example, one of the dangers of AI is that it can produce plausible but incorrect results. In construction, this could potentially result in flawed risk assessments or incorrect safety recommendations.

Supply chain

As recent cyber attacks on companies have demonstrated, greater use of AI and technology increases the vulnerability of supply chains to the risk of cybercrime. It can also have a significant impact on the resilience of the business.

Data collection, quality and security

Data management and data quality are fundamental to the success of AI. Poor data inputs lead to poor quality outputs. Two of the biggest hurdles to the adoption of AI are having trust in the data and obtaining reliable data in the first place. For AI to function effectively, data needs to be in a machine-readable format. However, as the construction industry has not been traditionally a “data-first” sector, data architecture in technology applications and software tools tend to be fragmented and inconsistent. Investment will therefore need to be focused initially on developing standardised and structured machine-readable data. As one executive put it: a lot of industry information is still only in people's heads, let alone in any database.

AI may inadvertently reinforce bias or operate outside legal and ethical boundaries, which means that there must be data quality assurance in place to check for the risk of biased data and inaccuracy. Over-reliance on AI may erode critical human skills and reduce oversight. In addition, many AI systems operate as a “black box” where users can see the inputs and outputs of the system but have limited understanding of what factors are used and weighted in reaching conclusions. This can undermine overall trust in AI technologies.

Data security, sharing of commercially sensitive information and protection of intellectual property could all be impacted by increased adoption of AI. Therefore, appropriate measures such as data-sharing protocols and data stewardship need to be put in place at the outset.

Opportunities for the sector

The construction industry is ripe for disruption. Projects take longer than planned, costs are often higher than anticipated, productivity has fallen because of higher levels of complexity, and there is a chronic shortage of skills. AI has the potential to improve productivity and create data analysis and insights to aid regulatory compliance and reporting.

Improving productivity

AI could help improve performance by creating efficiencies through automation and support better decision-making, including:

• optimising design and engineering processes by using generative design and complex simulations, including data layers on digital twins to enhance processes and predict potential problems and defects. Integrating AI with other technologies such as ‘internet of things’ (IoT) sensors – including smart appliances, autonomous safety systems and remote sensors that monitor manufacturing equipment – or building information modelling (BIM) tools – software programmes that create and manage digital building models – can create responsive, dynamic project environments, in which decision-making is based on real-time data rather than a static plan;
• exploring how contract management can be improved with fewer people and with better quality outputs;
• improving cost estimation and forecasting by using AI for predictive analysis – utilising data, statistics, modelling combined with machine learning to predict and plan projects;
• identifying and predicting faults and performance degradations in buildings and infrastructure;
• monitoring safety on site using drones and real-time tracking, as well as exploring how the use of AI and robotics could improve safety;
• improving data interpretation – AI analysis of lidar scans, photography and site surveys will enable better interpretation of data than simply relying on statistics and measurements.

Data monetisation

In the digital era, data is a valuable commodity. How do you value and measure efficiencies? Is the company’s investment in AI being passed on for free to clients? The concept of data monetisation is to generate tangible added value from data assets by either selling data or using it internally to improve operations, create new products or services, and enhance decision-making.

Regulatory compliance

AI can be used to improve sustainability in construction, such as through harnessing data to monitor energy consumption and to reduce waste through resource optimisation and better material forecasting, as well as for environmental, social and governance (ESG) reporting and compliance. AI can also be integrated into the "golden thread" concept in building safety to enhance digital monitoring throughout the whole lifecycle, particularly for higher-risk buildings, to ensure ongoing safety and regulatory compliance.

What should leaders do?

AI is widely expected to disrupt all parts of the global economy. The built environment sector will be no different.

Industry leaders will need to ensure that they are adequately briefed on the potential for AI and be fully cognisant of risks and opportunities of using the technology. Implementation of robust governance policies and agile AI strategies which can adapt to evolving technologies and market conditions will also be essential.

Harnessing skills and talent within organisations was identified as critical to the success of AI adoption, both in terms of recruiting digital engineers and upskilling the wider workforce. Experimentation and continuous learning through the creation of safe environments to trial AI, sharing both successes and failures, is also crucial.

Discussions highlighted the importance of data quality and the need for standardised machine-readable data at the outset of any project. Data quality is fundamental to successful good quality AI outputs. Therefore, continuous human oversight is required to reduce the risk of bias and check the accuracy of data.

There are five principal areas that business leaders should be considering:

• developing a clear vision of how AI might disrupt the way in built assets are developed and operated. This will also require a roadmap for change to be developed. Keeping this under review will be necessary as the pace of change accelerates;
• developing clear governance and risk management, particularly around data standards, ethics, and data-sharing protocols, will be essential. This should include consideration of cyber security, business resilience, and protection of intellectual property assets;
• standardised, machine readable and actionable data will be required with common data environments that will need to be shared on projects in the same way as other advanced manufacturing industries;
• people and talent will be crucial in all aspects of AI development and adoption. Attracting the right talent who will be engaged in modernising a traditional, conservative industry that is slow to change will be hugely important; and
• exploration and learning will be required to enable new ideas to be safely implemented. The built environment sector is a safety critical industry. It is difficult to trial new technologies and approaches on live projects. Creating environments for exploring the benefits of AI without impacting operational delivery will be essential.