AI adoption in architecture firms

Five surveys, five answers

Each of those figures is accurate for its own question. None of them is the adoption rate for the profession, because the profession has not been asked a single question. The table below sets out who was asked, what counted as use and when, which is the only way to compare them.

The comparison table

Why the answers run from 6% to 78%

Individuals or firms

AIA’s March 2025 report asked architects about their own work and found 6% using AI regularly. Its Firm Survey asked firms whether AI was used in day-to-day work and found a third saying yes. Both can be true at once. A firm counts as a user if anyone in it uses AI, so a firm of 60 people with three regular users is a yes on one survey and 5% on the other.

Regular use, any use, or experimenting

Deltek’s 78% is generative AI use, with no frequency attached. The Chaos and Architizer figure of 64% counts people who experiment with AI in daily workflows, which is a lower bar still, and it comes from a survey run with a rendering software company and published as sponsored content, so we label it rather than lean on it. AIA’s 6% is the strictest definition on the page: regular use for the job. The stricter the definition, the smaller the number.

Who is in the sample

Size explains most of the rest. AIA’s own 2025 report says adoption is driven by firms with 50 or more employees and by architects aged 50 or younger. The Firm Survey’s gap between 27% of small firms and 61% of large ones is the same pattern in numbers. Deltek’s sample is architecture and engineering firms that benchmark, which skews large, and its 70% should be read as the figure for that population. RIBA’s 59% covers UK member practices and is the only national figure outside the US on this page.

When

The dates run from 2023 data (AIA Firm Survey) to May 2026 (Deltek). Adoption moved quickly over that period in every series that has two points: RIBA from 41% to 59% in a year, Deltek from 53% to 70%. A 2023 figure and a 2026 figure measure different years as well as different populations, and this page lists both so that nobody quotes the older one as current.

What firms use AI for

The surveys agree more on use than on adoption. AIA’s 2025 report found chatbots, image generators and grammar and text tools the most common applications, and 84% of respondents optimistic about AI automating manual tasks. Deltek’s 78% figure for generative AI use points the same way. The concerns are as consistent as the uses: 90% of AIA’s respondents were concerned about inaccuracy, data misuse, security, authenticity and transparency, and 67% of RIBA’s believed AI increases the risk of their work being imitated, while 35% worried it threatens the profession. On the business case, 65% of RIBA respondents expected AI to improve construction productivity, 59% of Deltek’s said successful AI would create a significant competitive advantage, and 42% thought they could lose market share within two years without significant digital transformation, up from 33% a year earlier.

What no survey shows: any effect on fees

We looked for a study measuring whether AI has changed what architects charge, how they structure fees, or whether fixed-fee projects have become more profitable because of it. We found none. The closest evidence is Deltek’s 38% of firms reporting a measurable positive business impact, against 39% saying it is too early to tell, and that is a statement about impact in general, not about fees. A claim that AI will raise your fees is not supported by any published survey as of October 2026. What the evidence supports is a capacity claim: firms that use it report automating manual tasks, and a small firm can measure the hours that saves against its billable hours or its proposal turnaround. That measurement is the subject of a separate guide listed below.

What this means for a small practice

• The adoption gap by size, 27% against 61% in AIA’s 2023 data, means a small firm that uses AI well for proposals, visualisation and admin is still ahead of most of its peers. The gap will not stay open. RIBA’s and Deltek’s series both moved by more than fifteen points in a year.
• Measure before you claim. Only 38% of Deltek’s larger firms could point to a measurable impact. Time saved per proposal, days from enquiry to consultation, and hours per project stage are the three figures worth logging from the first week.
• AI also changes how clients find a firm, through answers that sit above the search results. Getting cited by AI search sets out what a small practice can do about that, and the AI search visibility page listed below covers how we approach it.

Method

Each figure names its survey, its population and its date in the sentence that uses it. Status labels follow our source list: Verified means we read the release or PDF, and UK practices and Sponsored content say what they say. We left out a published critique of the AIA 2025 report’s response rate because it comes from a single commentator; if AIA publishes its response rate we will add it. Figures were checked on 10 October 2026, and the date on this page changes only when a number changes. Update calendar: Deltek Clarity in May 2027, inferred from the 2026 release; the AIA Firm Survey every two years, with a 2026 edition possible late in 2026; and a 2026 RIBA AI report if one is published, none of which had surfaced by 10 October 2026. ScaleArc sells an AI-assisted growth system to architecture firms, so we have an interest in this subject. That is a reason to check the sources, which is why every one is listed below.