For finance
For delivery
Why AI in AEC is not delivering yet and what is actually in the way
- June 19, 2026
- Posted by: Fred Davidson
- Category: AI
Your firm is eighteen months into the delivery of a new regional hospital. Four disciplines, two specialist subconsultants, and a client managing a construction budget that has already been revised once. The structural package is complete. MEP coordination is ongoing. And somewhere in the middle of it all, a senior project manager is trying to work out whether the firm is still on the right side of its fee.
They cannot tell (not quickly, anyway). The timesheet data is in one system, the fee tracking is in another, and the last proper financial review was three weeks ago at month-end. By the time the picture is clear, the margin has already moved.
Next quarter, a similar project comes in. A civic center this time, of a similar scale and complexity. The estimate is built the same way it always is: experience, precedent, and a healthy amount of gut feel. The data from the hospital project includes the actual resource burn, the coordination hours that were never in the original fee, and the scope changes that eroded contingency; this sits in a folder nobody will open.
This is the starting point for any honest conversation about AI in AEC. The data exists. There is just no infrastructure connecting it in a way that the business can learn from. And until that mode of data management in architecture firms changes, AI has nothing solid to build on.
Exploring common AEC AI adoption barriers: The challenges and concerns holding firms back
The 2025 SPI Impact of AI in Professional Services report surveyed 146 firms across the Americas, EMEA, and APAC. Architecture and engineering firms came out with the lowest AI impact scores across virtually every business application in the study. PSA scores at 1.33. CRM at 1.47. Industry averages sit at 1.96 and 2.12 respectively.
These numbers reflect something specific about how AEC firms are structured, not how ambitious or technology-forward they are. Here is what is actually getting in the way.
Project data that never compounds
Every hospital, every transit hub, every civic building your firm delivers generates detailed operational data: fee performance against actuals, resource utilization by phase, scope change patterns, subcontractor cost variances. Most of it disappears into a project archive the moment practical completion is reached.
The result is that architecture firm data management stays ad hoc rather than systematic, and institutional knowledge stays locked with individuals rather than becoming something the business owns. When a similar project comes along, the estimate relies on the memory of whoever ran the last one. This is one of the most significant barriers to using AI in architecture and engineering at a business level: there is no accumulated operational foundation for it to work from.
No real-time view of margin across the portfolio
Ask a principal at most engineering consultancies how their portfolio is performing right now, today, and the honest answer is usually some version of “we will know at month-end.” That three to four week lag between work happening and financial visibility is where margin quietly erodes.
A project running 10% over on labor hours in week six is recoverable. The same overrun discovered in week fourteen, after month-end reconciliation, is a conversation nobody wants to have with the client. Predictive analytics in project management can surface these signals early, but only when the underlying data is connected and current. In most AEC firms, it is neither.
AI readiness and resource planning that runs on instinct
An engineering workflow across a multi-discipline practice involves constant resource trade-offs: who is available, who has the right expertise for this phase, where is the next capacity crunch coming from. In smaller firms this is manageable through conversation. As the project portfolio grows, it becomes genuinely difficult to hold that picture in your head.
The firms that manage this well are not necessarily smarter. They have better visibility. They know, in real time, where their people are committed and where headroom exists. Without that visibility, AI readiness remains theoretical, and resource planning stays reactive rather than strategic.
A workforce proficiency gap that runs deeper than training
AEC firms have among the lowest workforce AI proficiency scores of any segment in the SPI study. It would be easy to read this as a training problem, and training is part of it. But the more fundamental issue is that when operational data is fragmented and unreliable, employees never develop real confidence in AI outputs. They try a tool, the outputs do not reflect the complexity of their actual project environment, and they go back to doing it manually.
Proficiency grows when AI demonstrably helps. AI demonstrably helps when the data feeding it is trustworthy. These two aspects have to develop together, and right now most AEC firms are trying to build one without the other.
Client expectations that are moving faster than internal readiness
According to the SPI report, nearly half of clients across professional services already expect their partners to be using AI in delivery. In integrated project delivery environments, where owners, architects, engineers, and contractors share risk across the full project lifecycle, the pressure to demonstrate data-driven decision-making is becoming a factor in how firms are evaluated at the bid stage.
This creates a real tension. Firms are being asked to show AI capability externally before they have built the operational foundation internally. An honest AI readiness audit of most AEC practices would surface this gap clearly: the ambition is there, the infrastructure is not.
What this means in practice
AEC firms are not losing ground on AI because they are behind the curve. They are losing ground because the way work has always been organized, by project, by phase, by discipline, makes it genuinely hard to build the kind of connected operational picture that AI requires to be useful at a business level. This is the core challenge of AEC digital transformation: not technology selection, but data connectivity.
The good news is that this is a solvable problem, and the firms that solve it earliest will have a meaningful advantage. Not because AI will do their thinking for them, but because they will be making decisions with better information, faster, and with less of their margin disappearing between the cracks.
At sa.global, we work with architecture and engineering firms to build that foundation through evergreen. It brings project financial data, resource utilization, and delivery performance into a single connected picture without requiring firms to replace the systems they already use. Above it, empower provides the intelligence layer that translates that data into the kind of early signals and margin visibility that changes how decisions actually get made.
The 2025 SPI report goes deeper on where AEC firms stand across data maturity, workforce proficiency, and delivery performance, and what the highest-performing firms are doing to pull ahead. If you are starting to think seriously about what AI readiness looks like for your practice, it is a grounded and practical place to start. Get more insights here and download the report.
FAQs
What are the main barriers to AI adoption in AEC firms?
The main AEC AI adoption barriers are fragmented project data, poor real-time financial visibility, reactive resource planning, low workforce AI proficiency, and a gap between client expectations and internal readiness. Most firms have the data but lack the infrastructure to connect it in a way that AI can actually use.
How can architecture firms use AI in project management?
Architecture firms can use AI in project management to surface early margin warnings, improve resource allocation, and build more accurate project estimates from historical data. The prerequisite is connected architecture firm data management: timesheet, fee, and delivery data need to flow into a single operational picture before AI can add reliable value.
What is an AI readiness audit for engineering firms?
An AI readiness audit for engineering firms assesses how well the business can support AI-driven decision-making. It typically looks at data connectivity, workforce proficiency, and whether real-time financial and resource visibility is in place. Most audits find that the technology gap is smaller than the data infrastructure gap.
Why is predictive analytics important in construction project management?
Predictive analytics in project management lets firms identify cost and schedule risks weeks before they become problems. A labor overrun spotted in week six is recoverable. The same overrun found at month-end often is not. Early signals only work when project data is current and connected across systems.
Why are AEC firms behind on digital transformation?
AEC digital transformation is slower than other sectors because the industry organizes work by project, phase, and discipline, which fragments operational data by default. Institutional knowledge stays with individuals rather than compounding into business assets. Until that structure changes, AI tools have no reliable foundation to build on.
Fred Davidson
Fred Davidson is a seasoned marketing professional with deep expertise in demand generation, brand strategy, and revenue marketing for software services firms. With nearly twelve years of experience driving growth across global markets, he combines data-driven precision with creative storytelling to align marketing, sales, and delivery. Fred is passionate about building connected campaigns that accelerate pipeline, strengthen positioning, and deliver predictable growth.