ZGF Architects
PDX Reimagined: How a Digital Backbone Powered America’s Most Ambitious Airport Project
A new way to supercharge Revu with AI.
The Complete Guide to Construction Takeoffs in 2026
U.S. Edition
AEC Technology Outlook Report 2027
A survey of 750 architecture, engineering and construction decision-makers in the United States on what AI is doing in the field and the office, what it isn’t doing yet, and what separates the firms that got it working from the ones still trying.
1. Executive summary
- About 28% of U.S. AEC firms are running AI in production, up from 27% a year ago. Only 7% have AI embedded across multiple core workflows, and 32% are parked in pilots.
- Firms that got AI into production report real returns, so scaling AI is an operational problem.
- High implementation cost is the most-cited barrier to scaling AI, named by 50% of firms still piloting and 50% of firms that already scaled. Lack of internal AI skills is second, at 34%.
- Pilots are far more likely stuck on what to build and whether their data can support it — 37% vs. 20% on definitional problems, 15% vs. 6% on data readiness.
- AI has learned to predict, not to prevent. Cost predictability is the top benefit, at 64%; reduced rework sits near the bottom, at 17%. AI is concentrated in estimating, design and planning, and thins out closer to the jobsite (where rework happens).
- Proficiency with digital tools is the most sought-after hiring skill, at 93%, and the hardest to find, at 77%. The automation fix is gated behind the shortage it would solve.
- The industry has bought the ability to see problems sooner. It hasn’t built the path from seeing to stopping — through connecting disconnected systems, cleaning up data, and training people to trust the output.
2. Survey methodology
3. Where AI adoption stands
About 28% of U.S. firms are running AI in production, up barely from 27% a year ago. What grew instead is the middle of the market: 32% are piloting AI in limited use cases, and 31% are exploring it without having started. More than half the industry has started and not finished.
How we define “in production”
Throughout this report, “in production” means the combined share of organizations reporting either that AI is embedded across multiple core workflows or that it supports a few specific workflows. Firms piloting AI, exploring it or refusing it are counted separately.
The grouping is ours rather than a survey response option, and we apply it consistently in every section. Full question wording and base sizes appear in the methodology.
4. Cost, skills and the readiness gap
What separates firms that scaled AI from firms still piloting
Among firms that have moved AI beyond pilots, integration with existing systems and tools was named by 47% as the most important factor.
Cost doesn’t distinguish the two groups. High implementation costs were named by 50% of firms still piloting and 50% of firms that scaled. And firms still piloting are markedly more likely to be stuck on questions that come before implementation.
5. What AI does when it works
Greater cost predictability leads, named by 64% of firms using AI. Earlier identification of risks and issues follows at 58%, reduced administrative burden at 55%. All three are ways of knowing something sooner.
Rework is construction’s most expensive recurring failure — direct costs run about 5% of total construction cost, closer to 9% with indirect and schedule effects counted — yet it sits near the bottom of what AI currently delivers.
Why is AI not reducing rework?
Rework originates in decisions taken months earlier and surfaces on a jobsite, where the people who could stop it have the least access to the systems that saw it coming. It’s where AI sits that makes the difference.
Cost estimation leads at 69%. Design and modeling, 69%. Project planning, 63%. Document management, 60%. All of those happen before anyone breaks ground, or in an office while the work goes on elsewhere.
6. The workforce reset
Digital tools proficiency is the most sought-after hiring capability, named by 93% — a threshold rather than a differentiator. Project management follows at 82%, AI-related knowledge at 69%.
Is digital skill now worth more than trade experience?
Asked to weigh trade experience against digital adaptability, 39% lean digital, 25% lean traditional, and the largest single group — 35% — says the two matter equally. The data doesn’t say the industry has stopped valuing people who know how to build. It says that among firms forced to rank the two, more now rank the software higher.
How are labor shortages shaping AI adoption?
Shortages are close to universal — only 16% report none. The most common consequence is a heavier load on the people already there, at 74%. Counting every respondent, 82% say labor and skills shortages limit their ability to deliver projects to some degree.
A third of firms already using AI name a lack of internal AI skills as a scaling barrier, and AI-related knowledge is the second-hardest capability in the industry to hire.
What labor and skills shortages have done
| Effect | Share of affected firms |
|---|---|
| Increased workload for existing employees | 74% |
| Delayed project timelines | 49% |
| Increased project costs | 39% |
| Limited the number of projects taken on | 36% |
| Increased reliance on subcontractors/external partners | 33% |
| Reduced quality/increased rework needed | 20% |
| Increased safety risks | 15% |
Base: organizations reporting a shortage (United States, n=631).
7. Where project workflows breakdown
Roughly four in five firms say disconnected tools substantially hurt decision speed, workflow efficiency and risk exposure. Breakdowns occur most often at the design-to-construction handoff (41%), the office-to-field seam (41%), and in document management (38%).
Asked to locate their friction, firms point overwhelmingly at the places where work passes between groups of people, not between systems.
Are firms using AI to connect their tools or replace them?
Among firms using AI, 41% use it to connect and coordinate data across existing tools, and 37% to automate tasks within individual tools.
Consolidation has been the promised endpoint of construction software for a generation — one platform, one source of truth, one vendor to call. Firms are spending their AI budgets on the opposite.
How AI is being used in relation to existing tools
| Use of AI relative to existing tools | Share of AI users |
|---|---|
| To connect and coordinate data across existing tools | 41% |
| To automate tasks within individual tools | 37% |
| To replace multiple tools with a single platform | 23% |
Base: organizations currently using AI (United States, n=448). Single response.
8. What separated firms that scaled
Roughly a quarter of firms got AI into production, and 210 of them answered a question no one else was asked. (One caution: they’re describing their own success after the fact — the weakest evidence in this study.)
Integration with existing systems and tools leads at 47% — first-ranked in every market studied. The second tier is about people and information, not software: cross-team collaboration (29%), transparency of results (27%), high-quality data (26%), and employee training (26%).
Strong leadership support was named by just 11%, a clear ROI case by 12%, adequate vendor support by 10%. High implementation cost is the largest barrier for firms that haven’t scaled — yet among firms that did, a clear ROI case ranks near the bottom of what helped. Money stops people.
A business case doesn’t appear to be what starts them.
The returns are real: 40% of firms running AI in production report savings above $500,000, and another 31% between $100,000 and $500,000. (Self-reported estimate, base n=210. Not comparable year-over-year.)
What comes next
The firms that got through weren’t working with better models. They bought from the same vendors, in the same year, at similar prices. What they had was somewhere to plug it in.
The work that moves AI from pilot to production isn’t AI work. It’s the connective plumbing the industry has deferred for 30 years, and most pilots are stalled behind it. The industry now has technology that can tell it what’s about to go wrong.
Stop losing time to rework
See what production-ready looks like in practice.
Start Your Free TrialFor more on how AEC teams are closing the gap, visit the Built blog