It’s been a few years since AI tools started making their way into architectural workflows, and at this point, their presence isn’t exactly surprising. What began as experimentation has, for many, become part of the day-to-day, whether that’s generating early concepts, testing ideas, or speeding up visualization.
But where do things actually stand now? Has AI moved beyond quick wins and novelty, or is it still finding its footing in more complex parts of the design process?
To get a clearer picture, Chaos and Architizer ran their fourth global survey, gathering insights from nearly 800 architects and designers.
Below, we’ve pulled out some of the most interesting patterns from the data, with more detailed findings available in the full report.
Key findings
Looking at the survey results as a whole, a few clear patterns begin to emerge:
-
AI is most effective in early design, where speed and iteration matter most, with 43% of respondents identifying concept and pre-design as the area of greatest impact;
-
Satisfaction with AI remains mixed, pointing to a gap between what the technology can produce quickly and what architects need it to deliver reliably;
-
Time savings are real, but depend heavily on how AI is used, with 85% of users reporting efficiency gains, mostly in concept design and image-based workflows;
-
Architects are approaching AI thoughtfully rather than impulsively, adopting it where it proves useful while remaining cautious about its limitations.
Who took part in the survey?
The survey reached a broad mix of architecture and design professionals, with the majority (60%) coming from smaller practices of up to 19 people. Medium-sized firms (20–99 employees) made up 27% of respondents, while 13% represented larger firms with over 100 employees.
Geographically, half of all respondents were based in North America, followed by Europe (22%) and Asia (10%), with additional responses coming from across Africa, South America, and Oceania.
In terms of disciplines, most respondents (71%) worked in architectural design, alongside a significant share of interior designers (25%). The remaining participants represented planning, landscape architecture, and engineering.
Most survey respondents work in architecture (71%), followed by interior design (25%)
The state of AI in architectural design & visualization 2026
1. AI adoption is growing, but still experimental
At this point, it’s no longer about whether architects are using AI, but how far they’re willing to take it. Around 64% of respondents say they’ve already experimented with AI tools, yet only 20% report fully embracing them in their workflow. For many, AI still sits in that in-between space, something they test and revisit, but haven’t fully committed to, often due to integration challenges.
At the same time, momentum is clearly building. 74% of respondents say they’re likely to increase their use of AI in the near future. Even a small share of those currently avoiding it expect that to change.
So where is this growth actually heading, and what are firms planning to invest in next? The full report reveals which AI use cases are gaining the most traction, and where adoption is starting to move beyond experimentation.
74% are likely to increase their use of AI tools in architectural practice in the next 12 months.
2. Visualization and concept design lead AI use
AI has found its place in the earliest stages of design, where quick iteration matters more than precision. Most architects using or testing AI rely on it for image generation and enhancement, using it to explore ideas faster rather than finalize them.
Conceptual and pre-design phases stand out in particular, with 43% of respondents identifying them as the area where AI adds the most value. It fits naturally into this stage, where speed and flexibility are more important than accuracy.
What’s more interesting is where things might be heading next. The report points to a growing interest in using AI beyond early design, especially in more time-consuming parts of the process. Tools like AI-generated video are starting to gain traction, and there is increasing attention around applying AI to technical tasks like documentation.
This suggests that while AI is currently strongest in early design, its role may be expanding.
3. Lack of control and reliability limits AI’s potential
AI is delivering results, but not always the ones architects are hoping for. Most users fall somewhere in the middle, with 69% saying they are only “somewhat” satisfied with what AI produces. A smaller group reports being very satisfied, while relatively few describe their experience as outright negative. Overall, the picture is mixed rather than clearly positive or negative.
Much of this comes down to reliability. Nearly half of respondents (48%) cite inconsistent or poor output quality as their biggest challenge, and many say results still need to be checked and refined. At the same time, there are clear signs of progress. Around 70% feel AI-generated visuals reflect their intent to a reasonable degree, even if not perfectly, and 60% say it improves certain parts of their workflow, despite occasionally introducing new errors.
This leaves AI in a somewhat uneven position. It’s useful, but not always predictable. So what would make it more reliable, and where are the biggest gaps today? The full report explores where current tools fall short and what architects are looking for next.
Poor output quality or unreliable results are architect's biggest obstacle when trying to use AI tools in their design process
4. Time savings are real, but uneven
A large majority of users (85%) report that AI has helped save time in their workflow, whether through small efficiencies or more noticeable speed improvements. Most describe these gains as incremental, but a smaller group is seeing more significant shifts in how quickly they can produce work.
Unsurprisingly, these time savings are most visible in the same areas where AI is already widely used. Concept design and ideation lead the way, with 48% of respondents reporting the biggest gains there, followed by image-related tasks like enhancement. And in some cases, more advanced uses like material selection or asset generation are starting to emerge.
These gains don’t come without friction, however. For many firms, integrating AI into existing workflows remains a challenge, with 33% of users pointing to poor compatibility with their current software as a key obstacle. As a result, efficiency gains often depend less on the tools themselves and more on how well they fit into established processes.
This helps explain why time savings vary so widely. AI is clearly speeding things up, but its impact remains closely tied to its integration into everyday practice.
5. Architects are adopting AI carefully, not blindly
Adoption isn’t happening blindly. Even as AI becomes more common, many architects are approaching it with a level of caution that reflects how much is at stake. Among those not currently using AI, 61% say they feel pressure to adopt it, yet a notable share (35%) say they’ll only do so when it clearly adds value.
That hesitation isn’t driven by a single concern. Instead, it comes from a mix of factors, some practical, others more philosophical, that shape how and when architects are willing to bring AI into their workflow. What’s notable is that outright resistance remains rare, with most firms still somewhere in the process of evaluating where AI fits.
At the same time, attitudes are shifting. Concerns that once dominated the conversation appear to be softening, suggesting that familiarity with the technology is beginning to change how it’s perceived.
The full report unpacks these motivations in more detail, revealing what’s actually holding firms back, and what might push them to adopt AI more fully.
Overview
Taken together, the findings point to a profession that is steadily integrating AI, but doing so on its own terms. The technology is already proving useful, especially in early design and visualization, where speed and iteration can meaningfully improve workflows. At the same time, its limitations around control, reliability, and integration mean it hasn’t yet become a seamless part of everyday practice.
Rather than a rapid transformation, what emerges is a slower, more deliberate shift. As tools continue to evolve, the question is less about whether architects will use AI, and more about how it will fit into the way they already work.
📢 Dive into the results in a live webinar on April 16
To learn more about the survey insights, join Senior Director of Product Operations, Roderick Bates, for a live webinar on April 16.