In short
AI is already part of daily work in architecture and design, but much less so on the building site. According to the RIBA, 74% of UK architecture practices used AI in at least some projects in 2026, up from 59% in 2025. In construction, Eurostat counted 10.79% of EU firms using AI in 2025, the lowest of any sector.
Your design team can visualise a concept within minutes, while on site someone still sorts the inspection photos by hand. That gap between the design office and the site is the real story of AI in architecture, engineering and construction. Below you can read what the numbers say, where AI already saves time, what architects worry about and how a practice or builder can start. Independent measurements of how much time or money AI saves in building projects are still scarce. I think whether it takes hold in a company depends above all on the people who start working with it.
- In the RIBA survey of 2025, only 4% of those surveyed thought human creativity would no longer be needed for building design because of AI.
- In the UK, 13% of construction businesses used AI in June 2026, according to the Office for National Statistics.
- For some document types, England’s Extract tool does in 2 minutes what can take a planning officer up to 2 hours.
- The CITB forecasts that UK construction needs 750 extra architects a year.
- In the RIBA survey, 67% of architects worry that AI will increase the risk of their work being imitated.
Keynote by Richard
AI & the future of construction. The autonomous site, homes from the factory line and what it demands of tradespeople and leaders.
How many practices use AI?
A majority of UK architecture practices now use AI, and the share is rising fast.
The figures here count practices, not individual architects. The Royal Institute of British Architects (RIBA) surveys its members every year. In its AI Report 2026, 74% of architecture practices said they used AI in at least some of their projects, compared with 59% in 2025 and 41% in 2024. These are the RIBA’s own survey figures among its members, not a national statistic.
The official European figures point the same way. Eurostat does not publish a separate figure for architecture. Architecture and engineering offices fall under “professional, scientific and technical activities” in the EU’s NACE classification. In 2025, 40.43% of firms in that group used AI, the second highest of all sectors after information and communication (62.52%). The EU average was 19.95%. What do these firms use? In the same group, text mining was the most used AI technology (25.22% of firms), followed by technologies that generate pictures, videos or sound (18.72%) and natural language generation (17.74%). That fits the daily work of a design office: reading and writing long documents, and producing images of a design.
The biggest opportunities for AI in construction lie in project management, planning, procurement and digital twins, but also in controlling machines. AI is leaving the screen and becoming part of robots and other machines on the building site.
AI in architectural design
AI is mostly a fast assistant in the early design stages, not the designer.
Generative tools can produce many variants of a floor plan or facade, visualise a concept within minutes and test options against criteria such as daylight or energy use. We described how this works in practice in building the future: how generative AI is redefining architectural design.
Architects themselves are optimistic about the effect. In the RIBA survey of 2025, 65% think AI will have an increasingly positive effect on productivity in the construction industry. Another 65% feel it will help the sector meet its net-zero targets, and 63% believe it will enable architects and developers to create buildings that better meet current and future performance requirements. They do not expect AI to take over the creative work. Only 4% of those surveyed think human creativity will no longer be needed for building design because of AI.
Why does construction lag?
The closer you get to the building site, the lower the AI use, in Europe, the UK and the US.
The figures per region:
- EU.
According to Eurostat, 10.79% of construction firms used AI in 2025, the lowest of all sectors. Among construction firms that did not use AI, only 9.74% had ever considered it, again the lowest. - UK.
The Office for National Statistics reports that 38% of UK businesses with 10 or more employees used AI in September 2026, up from around 12% in late 2023. Around 60% of businesses in information and communication used AI; in June 2026 the share in construction was 13%. - US.
In an April 2026 working paper, researchers at the US Census Bureau found an AI use rate of 34% in professional, scientific and technical services, against a national rate of 18%. Sectors that rely on physical output, such as construction, have much lower adoption rates.
Why is that? Eurostat asked firms that had considered AI but did not use it. Across all sectors, 70.89% named a lack of relevant expertise. Next came a lack of clarity about the legal consequences (52.52%) and concerns about data protection and privacy (48.83%). These figures cover all sectors, not construction alone. There is also a practical reason. A design office works on screens, with drawings, models and documents that are already digital. A building project brings together a client, an architect, engineers, a main contractor and many subcontractors, each with their own systems. Data ends up spread over many parties. AI works best on data that is complete and in one place.
On the site itself, change is slower but real. Machines that dig or level on their own already exist, as we described in how autonomous equipment could push the construction sector into the future. Drones and image recognition are changing inspections, as set out in how technology is redefining the way we inspect physical infrastructure.
Where does AI save time?
In England, an AI tool does in 2 minutes what can take up to 2 hours on some planning records, with people checking the result.
One of the clearest public examples comes from England, where the government uses AI to speed up planning. The Ministry of Housing, Communities and Local Government built a tool called Extract with the government’s Incubator for Artificial Intelligence. It turns old planning documents, such as conservation areas and tree preservation orders, into structured data. Officers can spend up to 2 hours manually working through one of these records. For some document types, Extract does the same work in 2 minutes. Since June 2026, every local planning authority in England has free access to it. Two details matter. First, people stay in charge: users review everything Extract produces and correct it before export. In around two thirds of cases, only minor edits were needed. Second, the tool works on data, which is exactly what planning departments, like many building projects, lack.
A second tool is still a prototype. According to the UK government, it aims to halve the time it takes to process householder planning applications, from 8 to 4 weeks in an average case. It is being tested with Barnet, Camden and Dorset councils. The government writes that every assessment will be reviewed and approved by a qualified planning officer before any decision is made. Whether the tool reaches its aim is not yet known. For architects, this matters directly: faster and more predictable planning decisions could shorten the time between design and construction.
Will AI replace the architect?
Construction still needs more people, but most practices expect AI to reduce staff.
The UK’s Construction Industry Training Board (CITB) forecasts in its Construction Workforce Outlook 2026-2030 that the industry needs an average of 41,200 extra workers per year. In the CITB’s figures, 37,740 architects worked in UK construction in 2025, and 750 extra architects are needed each year. Relative to their numbers, civil engineers (2.2%) and architects (2.0%) are among the occupations with the strongest demand.
Architects are not unworried, though. In the RIBA survey of 2026, 59% of practices believe that AI adoption will lead to staff reductions across the profession, and 61% agree that it will make it more difficult for early-career professionals to gain essential skills and experience. That is an expectation, not a measured loss of jobs. A year earlier, in the survey of 2025, only 18% expected AI to lead to job losses, while 35% were concerned that AI would threaten the profession.
Risks of AI in architecture
Architects worry most about imitation, design without expertise and the reliability of data.
The RIBA survey names the main concerns of architects themselves:
- Imitation.
67% worry that AI will increase the risk of their work being imitated. - Design without expertise.
44% are concerned that AI will enable people without sufficient professional knowledge to design buildings. - Reliability.
The RIBA also points to concerns about the reliability of data.
The common answer is the one the English planning tools already use: AI proposes, a qualified person checks and decides. For a building, that is not a formality. Errors in structure, fire safety or ventilation are expensive to fix once something is built.
What rules apply to your firm?
The AI Act already applies to firms that use AI, and responsibility stays with the professional.
In the EU, the AI Act already applies to firms that use AI, not just to firms that build it. According to the European Commission, Article 4 on AI literacy has applied since 2 February 2025: organisations must take measures to support the AI literacy of their staff. National authorities supervise this from 2 August 2026. No certificate is needed; organisations can keep an internal record of trainings. Data protection rules keep applying too, for example when site photos or camera images show people. And a tool does not replace professional judgement: have the architect, engineer or contractor check what it finds and what it may miss.
How do you start with AI?
Start small, with one task that comes back every week, and keep a person responsible.
- Pick one task that comes back every week.
Think of writing a design statement, summarising a tender, sorting site photos or checking a planning document. Not “AI across the whole firm”. Then walk through that process step by step, from the design statement to the site inspection: which tasks can AI take on, and which belong with your architects, engineers and site staff? - Start where your data is already digital.
Drawings, models, specifications and photos that are already in one place are the easiest start. That is one reason design offices are ahead. - Test on real projects.
A tool that works in a demo does not automatically work on your project. Compare its output with what your own people find. - Keep a person responsible.
AI suggests and points out. The decision on quality, safety and design stays with the professional. - Agree on what you feed it.
With 67% of architects worried about imitation, decide which drawings and client data may go into which tools. - Train your people.
In the EU it is required, and it is also the answer to the barrier most firms name: a lack of expertise.
More applications, from scheduling to safety, are covered in five ways generative AI is reshaping construction. AI and other new technology, from digital twins to machines that dig on their own, call for a different attitude in a trade built on craftsmanship. I recommend playing and experimenting with it: small, on a real bottleneck. That is the quickest way to learn how it works. That takes managers who give their people that room and then look critically, together, at what works and what does not. That side of change is also the subject of the keynote on the organisation of the future.
Richard speaks on this topic
AI & the future of construction. The autonomous site, homes from the factory line and what it demands of tradespeople and leaders.
Frequently asked questions
How is AI used in architecture?
Mostly in the early design stages and in documents: generating and visualising variants, testing options against daylight or energy, and reading and writing long texts. In Eurostat’s figures, text mining and image generation are the most used AI technologies in the sector that includes architecture and engineering offices.
How many architecture practices use AI?
In the UK, 74% of architecture practices used AI in at least some of their projects in 2026, according to the RIBA AI Report 2026, up from 59% in 2025. In the EU, 40.43% of firms in professional, scientific and technical activities, which includes architecture and engineering, used AI in 2025, according to Eurostat.
Why does construction use less AI?
Construction work is physical and spread over many parties, so data is often incomplete or scattered. In the EU, 10.79% of construction firms used AI in 2025, and in the UK 13% in June 2026. The most common barrier Eurostat measured across all sectors is a lack of expertise.
Will AI replace architects?
The figures do not show that yet. The CITB forecasts that UK construction needs 750 extra architects a year, and only 4% of architects in the 2025 RIBA survey think human creativity will no longer be needed for building design. Still, 59% of practices in the 2026 RIBA survey expect AI to lead to staff reductions across the profession. The work itself will change, with more time for design and less for routine tasks.
Where should you start?
Pick one recurring task, agree who checks the result and measure what it saves.
Choose one recurring task in your office or on your projects, write down what AI may do and who checks the result, and measure after a month how much time it saved. The design office has shown that AI can be part of normal work. Whether the building site follows, and how fast, is still open. AI in architecture and construction is one of the themes in the keynote on the AI revolution.
Richard van Hooijdonk is a trend watcher, futurist and international keynote speaker. Many trend watchers and futurists look at the future from a distance. Richard van Hooijdonk lives in it. His team consists of people and AI agents, and he has several chips implanted in his body. He has given more than 3,000 keynotes in more than 30 countries. View his keynotes →