Our third white paper from the Future Business of Architecture programme.
A source simultaneously of excitement and apprehension, AI holds the promise of enhancing design capabilities and streamlining processes, while raising complex issues about authorship, ethics, and the future of creative jobs. For the Future Business of Architecture programme – which sets out to take a research-based 10-year view of practice – technological innovation, fuelled by AI, represents the greatest unknown as a potential force of both disruption and positive change.
AI is rapidly growing in its sophistication, applicability, and ease of use. Even in the first half of 2025, the improvement and availability of consumer language models, such as Google Gemini 2.0 Flash, OpenAI o1 and Microsoft 365 Copilot, have led to widespread adoption at home and at work. Now embedded in search and video conferencing software, AI has transitioned from a much talked about technology to a commonplace tool on our screens and in our phones. In an architectural context, AI follows previous waves of technological innovation, from the printing press, to computer-aided design (CAD), to building information modelling (BIM) and digital twins, all of which have had significant, sometimes radical, effects on the practice and business of architecture.
How rapidly AI will continue to develop and how it will usefully and profitably be applied in practice is very unclear. It is a topic of much current and useful discussion, debate, and discovery, within and outside the profession. This makes the task of looking ahead 10 years (when who knows what the next six months will bring) a difficult one. Future scenarios are uncertain. Possible futures vary wildly in both what technologies may be developed and what effects their use may have on jobs, trades, professions, and wider society.
With that in mind, this paper restricts itself to an overview of three possible outcomes (the good, the bad, and the little changed), what effects they may have on the business of architecture, and what we learnt from our research about architects’ current views on AI and technology. But before we explore possible future scenarios, the starting point is establishing the status quo.
UK architecture: a leader in technological innovation
The profession is positive about the business benefits that technological innovation can bring. In our Future Business of Architecture Survey, two-thirds (67%) of respondents thought that digitisation and technological innovation would have a positive effect on their future business.
The UK has a strong record in technological innovation, in leading the digitisation of construction. Technological innovation is not something that has passively happened to UK architecture, but something UK architecture has led on and shaped.
With the introduction of CAD and early BIM systems from the 1970’s onwards, architecture began its shift from paper (and sometimes wood, plaster, or paint) being the primary means of representing design intent towards whole project digitization, through the digital representation of drawings and models.
The UK government then took a lead with the 2011 Construction Strategy, and the ‘BIM Mandate’, which required all centrally procured public sector projects to use BIM Level 2 by 2016. This created a reliable demand for BIM, so accelerated its integration into practice. The NBS reported that BIM adoption among construction professionals in the UK increased from 13% in 2013 to around 70% from 2018 onwards.
Successful project collaboration using BIM requires project information to be in an agreed format, so that it can reliably be read, modified, shared, and organised. Standardisation is needed. Here again, the UK took the lead. The international standard for BIM, the ISO 19650 series, arose directly from the UK’s BS 1192 series.
The requirements of the golden thread of information and the Building Safety Act 2022 suggest that the UK’s leadership in standardised information remains.
So, through CAD, and then BIM, UK architects have demonstrated an ability to adopt new technology and adapt it to their needs, and so transform practice to collaboratively create better buildings.
Indeed, many think of BIM as a stepping stone towards AI – that holding design data in a standardised, structured way is a precondition for the successful deployment of AI tools in a practice. Or, to put it more simply, without good training data there can be no AI. If this is right, UK architects are among those best placed to successfully implement AI in practice.
AI adoption
Within the profession, AI is already being used for a range of tasks, including early-stage visualisations, specification writing, compliance checking, and product selection and analysis. AI is also helping practices with the day-to-day tasks of practice management, including report writing, bid creation, client management, and project scheduling.
It is also a time of exploration, with practices and individuals pushing AI to integrate with an ever-expanding range of design and management tasks, to improve existing processes, and to do wholly new things. A lot of novel thinking about AI and the profession, and ways of working with it, are being devised, explored, shared, and developed. There’s a feeling of rapid innovation in the air.
AI and the future
The 2025 RIBA Artificial Intelligence Report suggests that AI adoption is rapidly increasing among architects:
"More respondents than last year reported their practices are using AI in the projects they are working on, and are using it more often and to do more things. In 2025, 59% of practices reported using AI for at least the occasional project, up from 41% in 2024. Most practices are now using AI. Conversely, the proportion of practices that never use AI has dropped, from 59% to 41%."
Source: RIBA Artificial Intelligence Report 2025.
In the near term, AI adoption looks set to gather pace within the profession.
"Architects expect AI to be used in more areas than currently, improving efficiency and accuracy in the design process and becoming more integrated into project management.
A majority of respondents (53%) expect their practice to have an AI policy within the next two years, and nearly as many (47%) anticipate their practice investing in AI research and development."
Source: RIBA Artificial Intelligence Report 2025.
Looking ahead to the next 10 years, the Future Business of Architecture Survey suggests that AI use is the activity that is going to grow in importance over the coming decade. Eighty-eight per cent of respondents thought it would become more important – ahead of business development (61%) and client engagement (51%). Other technological activities are also seen by most practices as set to become more important, including BIM creation and management and data analysis, suggesting that AI will develop from and require BIM and data analysis, rather than replace them.
Technology and importance to future business Source: RIBA Future Business of Architecture Survey 2025
Drilling down further, the Future Business of Architecture Survey asked respondents about how far each of the RIBA Plan of Work stages might be affected by AI in the coming 10 years.
The scale was from ‘transformational’ (where the stage would be fully automated) to ‘no effect’. The findings suggest that, presently, architects are more likely to think that AI will have minimal effect or no effect at all than it is to be transformational. Fewer than 10% think that any of the work stages will become fully automated. That said, for every work stage, the majority think the future will be somewhere in the middle – that AI will have a significant or moderate effect.
AI and the RIBA Plan of Work stages. Source: RIBA Future Business of Architecture Survey 2025.
In more detail, Stages 2 to 4 are expected to be most affected, with over three-quarters (78%) of respondents predicting AI will have a transformational, significant or moderate impact, with Stage 2 most likely to be thought as becoming fully automated. Stages 0 and 1 and the construction stage are also expected to be significantly affected by AI. The handover stage is seen as being least affected, with nearly half expecting minimal or no effect.
Predictions vary, but it looks like AI will be increasingly adopted and will become more important to the business of architecture in the next 10 years.
[Midjourney image here]
 |
 |
Will Garner uses Midjourney to explore design pathways. Published in Image Generation: Artificial Intelligence, Creativity and Design (RIBA Publishing, 2025), edited by Hamza Shaikh
Learning from the past - BIM and digital transformation
When BIM was in its infancy, the promises made for it were great. There would be a single digital model of a building, born in the early design stage, growing as design and construction progressed, and with a life as long as that of the building itself – a model that would contain all the required project information in a reliable, standardised format, so promoting collaboration, improving efficiency, and reducing error, waste, and rework. The complex, risk-passing, rivalrous, and disputatious construction sector was set to be transformed.
"BIM’s early promise was prodigious. By representing the physical and intrinsic properties of buildings with a software-based digital rendering, these specifications could be readily shared between stakeholders during the building’s lifecycle. This would usher in an era of real-time collaboration not only for design, but also for construction and operation.
Thirty years later, however, and supporters’ promises that BIM would revolutionise the industry are mostly unrealised. While BIM now produces excellent drawings, the full vision for it outpaced its underlying technology."
Source: RIBA Horizons 2034, Jesse Devitte, Innovation strategy.
And yet, while BIM has not delivered on the promises of its early evangelists, it has delivered very real benefits and efficiencies to the design and construction industries. In the UK there are exemplar practices and projects that demonstrate the huge value BIM can bring. The golden thread of information remains fundamental to improving the sector’s performance, to creating safe and sustainable buildings. The early days of innovation can bring over-promise, but in the end, real benefits are brought.
With that note of caution, let us turn to three scenarios for the future of the business of architecture in light of AI and technological innovation, beginning with a bleak scenario, but one that is possible.
"Certainly, you can see AI significantly changing the way that we work in the future … I think there's two sides to that. Firstly, it's about saying, okay, well, how might AI better support us in the way that we work. So we can do any manner of different things a little bit more quickly. We can do things more efficiently. But then, equally, what can AI straight up replace? At the moment, it still needs a filter of quality and understanding that can only come through the experience and knowledge of a human being, of an architect. It's when that point begins to shift, and I don't know when that will be."
Source: RIBA Future Business of Architecture Qualitative Research, 2025, large practice.
Scenario one - the bad future
"We will neither need nor want doctors, teachers, accountants, architects, the clergy, consultants, lawyers, and many others, to work as they did in the 20th century."
Richard Susskind and Daniel Susskind: ‘The Future of the Professions: How Technology Will Transform the Work of Human Experts‘
This scenario imagines what might happen if the impact of AI is as great as that of the Industrial Revolution and what implications this could have on the practice of architecture.
For society, the coming decade sees a rapid increase in inequality. People are replaced by AI, leading to endemic unemployment and the hollowing out of the middle classes. The world’s wealth becomes ever more concentrated among a handful of owners of tech companies: the owners of the AI systems and algorithms. The climate emergency accelerates, as fossil fuels are burnt at an increasing rate to power AI data centres.
AI, as a superior intelligence, begins to treat people very much like we have treated other species of lesser intelligence – an exploitable resource.
At the end of the decade, we are heading into a future in which there are no uniquely human applications of intelligence. Creativity, design, spatial reasoning, ethical thinking: these can all be done as well or better by machines and done faster at lower cost. There are no intellectual monopolies left for people.
The ‘grand bargain’ of the professions has effectively broken down. The social contract through which professions were given exclusive rights to deliver specialised services has been broken, as digital systems make profession expertise readily accessible and inexpensive.
In this worst plausible case for the future of architecture, the profession is significantly reduced in its size and importance, as architects, like many other professionals, fail to find a role post-AI.
The range of activities that comprise the practice of architecture have been broken down into a series of AI-system executable tasks. AI simulates empathy and design dialogue, replacing architects in the briefing stages. Design has become algorithmic, and AI systems that have trained on vast datasets of buildings can generate unimprovable designs in a trice, designs that precisely and demonstrably fulfil client needs. Complying to regulation is automated, as planning submissions and building regulation compliance checks are carried out by AI. Construction is largely automated, with AI spanning both design and material selection and creation. Post-occupancy evaluation is done in real time, by building sensors linked to an always-up-to-date digital twin. Building residents are similarly monitored, with buildings adapted, often in real time, to induce client-desired behaviours from their occupants.
The systems that carry out these tasks are available to contractors and clients, just as they are to architects.
Ultimately, in this dystopian, technology-dominated future, architects, like the other professions, are marginalised. Their role is reduced to keeping a watching-eye on an AI system, perhaps accepting some legal and ethical responsibility, and stepping in to manage unexpected outputs.
Architects and craft
Even here, however, a distinct – and potentially profitable – role may still exist for architects.
As described by Veblen, goods and services are not always purchased because of how well, or for how long, they fulfil a function. Instead, many goods and services are purchased because they denote the social status or group-belonging of the purchaser. Leisure activities, designer clothes, some makes of cars, or even certain books or magazines, can be purchased as a social status or group-belonging identifier. This can be bound up with the notion of exclusivity. Indeed, the higher the price the better – it denotes greater exclusivity.
Should AI enable the rapid, reliable, and comparatively cheap production of buildings and building design, a market may well remain for human-crafted architecture. Not necessarily because it is any better in fulfilling the function of the building, but because it is exclusive and available to some, but not all. This may be as true for commercial clients as domestic customers, and could include imaginative renovations and retrofits as well as new-build projects. Perhaps we will see the wider return of architects branding their buildings, and high fees as a brand differentiator.
This leads to another avenue for future business prosperity. Following the Industrial Revolution there was a widespread intellectual shift towards the Romantic, Gothic, and Arcadian. Craft was valued above machine production, the human crafted object having an intrinsic value, not replicable by machine, even though the crafted object itself could be replicated. The Arts and Crafts movement and the Pre-Raphaelites emerged. This found its expression in architecture too, advocated strongly and at length by Ruskin.
So, just as a copy of a work of art or craft is not as valued as the original, perhaps even in this dystopian future architecture as a (master) craft for wealthy business and people will remain. In a world where wealth abounds, albeit concentrated, this may be a lucrative market.
Scenario two - the future is much like today
This leads us a second possible future for architecture and technology. This scenario assumes that we are currently experiencing a period of AI hype, and that progress over the next 10 years is actually slight. Things stay broadly as they are, albeit with some useful improvements.
While AI applications develop over the next 10 years, the growth of their capabilities slows significantly, as large language models fail to develop meaningfully beyond their current capabilities. AI adoption is slowed by resistance from those currently holding the jobs that might be displaced. AI applications continue to be useful, but they remain prone to error and inaccuracy, and their propensity to put user-pleasing content above accuracy remains. Attempts to replace human activity with AI fails, except in all but the most trivial or specialist activities. General intelligence fails to materialise. The imagined plethora of real-world business case applications – those with levels of acceptable risk and return on investment – also fail to materialise. AI providers are unable to sustain current levels of investment, without sufficiently attractive applications. The energy use of AI proves unacceptable to a world increasingly grappling with the urgent and crippling effects of climate breakdown.
In the profession, these elements work together to make AI and its application very much like it is today, but a bit improved. It is an increasingly useful tool for architects, assisting with design visualisations, early-stage design iterations, specification writing, report writing, and practice management. Perhaps more junior roles will be displaced, but the fundamentals of the profession remain the same. Design expertise and critical, spatial thinking cannot be learnt and applied by machines. The ethics of architecture – of understanding and responding to the sometimes-conflicting needs of society, clients, contractors, the delivery team, and even the AI system – remains an area for human professionals and not AI. Liability for design decisions needs a resting place, and it stays with the designer, not the tool. Design thinking and human creativity, it turns out, remain irreplaceable.
 |
 |
Arka Works is using AI to bring early design sketches to life: a simple digital drawing (left) can quickly be turned into a detailed, realistic image of a neighbourhood (right). Published in Image Generation: Artificial Intelligence, Creativity and Design (RIBA Publishing, 2025), edited by Hamza Shaikh.
Scenario three - the good future
In this final scenario, the immense technical promise of AI is largely realised. Its capabilities rapidly develop and machine-based intelligence (whether general or more task specific) becomes widely affordable and available. The change in the economy and society is again comparable in magnitude to the Industrial Revolution, but it happens within a decade or two. Let’s also assume the effects are largely benign – good for society, good for the professions. What might such a future look like?
Up to now, intelligence has been limited by human capability, capacity, population, discipline and attention. But over the next 10 years, intelligence becomes effectively unlimited, and we see a huge rise in innovation. Many of our most challenging problems are resolved through AI. New medicines, materials, modes of transport, means of production, and power sources are now in reach, to be delivered within the next 10 years. Global wealth has increased exponentially, along with health and wellbeing.
To ensure a stable society in an era of plenty, wealth is well distributed, or at least sufficiently distributed to allow all to live with sufficiency and dignity. Although jobs have been displaced, other, more fulfilling roles take their place. Humanity is on the brink of resolving the climate emergency, producing limitless energy, ending poverty, and seeing average life spans increase by years or even decades.
In this benign future, we can see an enhanced and growing role for the architects’ profession. This has been well described by Mark Greaves in the RIBA Horizons programme:
‘Recently completed projects are noticeably safer, more harmonious and more sustainably constructed. The use of AI throughout architectural firms has enabled an explosion of design creativity, coupled with a more collegial relationship with engineers and builders. Even the smallest and most routine structures bear the touches of thoughtful design.
Clients are delighted, professional employment is stable, and fees exhibit steady growth commensurate with the greater overall value that the profession brings.’
Source: RIBA Horizons 2034, Mark Greaves, Architecture in the age of AI.
Helpfully, Greaves provides four signposts to indicate that we are on the way towards the ideal future in architecture:
Acquiring professional knowledge
AI systems absorb the specialised knowledge, abstractions, and epistemology of architecture, through being trained on the vast quantity of publicly accessible information about architecture. For practices deploying AI, their AI tools will integrate private data and information so that ‘practice knowledge’ can be gained.
Achieving human-like judgement
AI systems become human-like in their ability to make professional, context-sensitive judgements based on incomplete and imperfect data. These judgements are not just about building structure or fabric, but encompass ethical, cultural, and perceptual considerations too. They are explained to the client, who will commission projects based on the judgements and decisions of AI. AI is a long way from this now, but it is a plausible scenario for 10 years’ time.
Integrating into business
AI speeds up many tasks and makes practice more efficient. Examples include rendering, document production, compliance checking, and exploring the implications of design choices. While increasing efficiency, this integration may reduce the cost of providing traditional architectural services and displace roles, allowing practices to reduce their fees. At the same time, new roles may emerge to exploit AI’s potential. The source of fees may similarly be transferred from the now cheaper-to-produce services to new, more expensive AI-enabled services.
Clarifying professional responsibility
The relationship between professional responsibility and intelligent machines will need to be made clear. Part of being a professional is taking responsibility for professional decisions – and accepting the attendant liability. As AI becomes more capable, the profession must define who is accountable for decisions made with or by AI and legal frameworks will be required to establish the levels of responsibility AI can bear. Will legal and professional responsibility be delegated to intelligent software?
 |
 |
An AI-generated model alongside an AI-generated design for a building by Carlos Bañón, illustrating the potential disruption to traditional architectural workflows. Published in Image Generation: Artificial Intelligence, Creativity and Design (RIBA Publishing, 2025), edited by Hamza Shaikh.
A new approach to fees
Over recent years, practice profitability has become increasingly challenging because fees are under pressure and costs are on the up.
According to the most recent data from the Architects Council of Europe, most practices in the UK use one of two ways to calculate their fees: either as a percentage of contract value (40% of current jobs) or as a lump sum (40%). A further 15% of current jobs are charged at an hourly rate. There is some variation across Europe, with hourly rate charging being more common in the Nordic countries – which often enjoy relatively high hourly charge-out rates.
Nevertheless, the existing methods of charging for an architect’s work do not reflect the value that good design brings to a building’s performance. As AI accelerates, the sector may move towards a performance- or outcome-based model for fees.
Under this scenario (the good future), fees will be based on how well a building fulfils its individual and environment-wide functions, rather than on the cost of constructing it. Much is already possible. But AI will increasingly offer the opportunity to measure the whole-life outcomes of a building at the design stages. The real-world value of good design can be modelled, quantified, and given an accurate financial value. Fees can, in principle, then be charged based on the demonstrable added value brought by an architect.
For all buildings, this added value could be the effect they have on the health and wellbeing of people and on the environment. As the climate emergency gathers pace, it can also mean designing buildings that are sufficiently resilient to be insurable for the long-term. For specific buildings, it can mean the how well they fulfil their client-required function. For hospitals, this may be how much more quickly and fully people are made well; for schools, it could be improvements in pupil achievement and development; for a retail vendor, it may be increases in footfall and revenue received per visitor.
Another business model, servitisation, offers a future where ‘the buyer does not purchase the asset but leases its availability, paying for delivery of its service at the agreed level rather than making capital payment’. Just as Rolls Royce jet engines are leased for the power they provide, or lighting installations for the illumination they give, so too can buildings be leased on the basis of their performance. And again, in this future, architects will be able to charge based on the demonstrable improvements in building performance that their designs bring.
So, in this good future, not only is the future business of architecture preserved and enhanced, but revenue becomes commensurate with value added. Good design brings in good money. Bad design does not.
Signposting the future of AI
No single future path ever exists in isolation. It is most likely that the future will be a mishmash of all the above scenarios.
For practices of any size, failing to maintain relevance and invest in upskilling may lead to diminishing prospects in an increasingly competitive landscape.
Smaller practices may find themselves particularly vulnerable, as AI tools for standardised domestic design become accessible to non-professionals, enabling the generation of basic plans and concepts without professional oversight. Larger practices focused primarily on early-stage design, without responsibility for delivery, may also face disruption.
The adoption of new technologies often accelerates polarisation. Some smaller firms may choose to differentiate by targeting the high-end, bespoke market, serving affluent clients and investing in their design identity through aspirational content on platforms like YouTube and social media.
For most practices, however, AI adoption is likely to be piecemeal rather than strategic, driven by convenience and affordability. Tools will be selected ad hoc to save time on both architectural work and running a business. Much AI functionality will be embedded in everyday tools.
In contrast, the largest global firms are well positioned to invest in AI, using proprietary data to enhance design quality and process efficiency, and demonstrating value through the measurable performance of their buildings.
As the bad future scenario illustrates, there are risks to the profession from AI. The technology is likely to disrupt established practice and may displace professional roles, enable imitation, and allow individuals lacking appropriate competence to offer building design services. The legal and ethical implications of AI are significant and complex.
An awareness of AI-related risks may renew an emphasis on human values: meaningful client engagement, ethical integrity, professional accountability, originality in design, and a deep understanding of place and context.
There will always be a role for the profession.
Adrian Malleson, Head of Economic Research and Analysis, and Helen Castle, Director of Publishing and Learning Content at RIBA, are co-leading on the Future Business of Architecture research programme. Adrian has recently also led on RIBA’s Artificial Intelligence Report 2025.
The Future Business of Architecture is a RIBA Horizons Programme, sponsored by Autodesk.
