Reach a highly engaged audience of AEC innovation leaders, executive teams, and industry titans. Check out our media kit.

The $12 trillion industry may have finally found its translator 

For decades, the construction industry has chased a single idea: if every stakeholder adopted one universal digital language, the sector could finally modernize. That vision consumed enormous energy and produced limited results. AEC remains among the least digitized industries on earth, still reliant on drawings, PDFs and emails. 

A different path has quietly opened. Speaking on the Bricks and Bytes podcast, Alain Waha, Chief Technological Officer of global engineering firm Buro Happold, shared a compelling perspective.  

According to Alain, this moment could be the most significant technological inflection point the built environment has seen since steel and electricity. His argument is straightforward: rather than forcing the industry to conform to machines, a new generation of AI can meet it where it already is. 

The AEC industry has long described its fragmentation through the metaphor of the Tower of Babel. Thousands of firms, each working in different formats and standards, unable to speak a common digital language. The traditional answer was a master ontology, a universal data schema, and an industry-wide ERP. None of it scaled. With 90% of construction firms employing fewer than 50 people, the economics of enterprise integration never worked. 

Alain reframes the question. Instead of teaching everyone Esperanto, why not build Google Translate? 

"We had this vision that the way humans build needed to be moved into the way computers think," he said. "And now we've actually created technology that approximates how humans can think and read." 

In practice, that means large language models can now read a drawing, cross-reference it against a contract, and flag discrepancies. They can locate the design report, base of the design, and analytical model for a particular structural system, a task that might take an engineer several hours. The technology works with the industry's existing information rather than requiring its wholesale replacement. 

For a sector told for a generation that it must digitize before it can improve, this is a genuinely structural shift. The barrier to adoption always hinged on the fact that the technology demanded too much change from too fragmented an industry. 

The commercial implications are already creating anxiety. If an engineering deliverable previously priced on human hours can now be produced in a fraction of the time, the cost-plus model that underpins most AEC fee structures faces direct pressure. Alain points to the volatility in share prices of major engineering consultancies as evidence that the market is already pricing in this future. 

But cost erosion is only the destination for firms that define their value in terms of time spent. 

The alternative is to define value by outcomes. A well-engineered building produces better occupant experiences, lower lifecycle costs, and stronger environmental performance. These are outcomes clients will pay for, independent of hours consumed. 

Alain is candid about the current state of affairs. London alone faces a shortage of roughly a million homes. Net zero construction at an affordable cost remains unsolved. Projects are chronically late and over budget.  

If AI is used merely to make the same mediocre outcomes cheaper, the sector will have squandered its opportunity to innovate and improve the quality of construction projects. 

One of the more nuanced points in the conversation concerns what happens after AI saves an engineer four hours. The instinct is to do more. Alain warns against this. 

A probabilistic system produces the most likely correct answer, not a guaranteed one. The saved time should be spent reviewing, questioning, and refining the output, not increasing throughput. 

He draws an analogy to navigation. An experienced sailor instinctively knows where north is. A generation raised on GPS does not. If AI handles the routine cognitive work of engineering, how do junior professionals develop the judgment to assess whether the output is sound? The risk is not that AI replaces engineers but that traditional engineers lose the capacity to evaluate what AI produces. 

There is also institutional knowledge. The tacit expertise of experienced professionals has always been the core asset of AEC firms. When those people leave, the knowledge leaves with them. 

Alain sees knowledge capture as a strategic priority, enabled by the same AI tools that augment daily work. Every project completed should feed a knowledge system that makes the next project better. Firms that build this effectively will compound their advantage. Those that do not will lose it, one retirement at a time. 

Current LLMs excel at processing text: contracts, reports, specifications. But they do not understand three-dimensional space. They cannot reason about floor plates, structural loads, or the physics of how a wall responds to force. Concepts that a three-year-old grasps intuitively remain beyond the reach of frontier AI. 

The capital is starting to flow. Alain points to Autodesk's recent $200 million investment in Feili's spatial AI start-up and emerging work in world models inspired by researchers like Yann LeCun.  

But building foundational models for the built environment requires training data that the AEC sector alone may struggle to generate, particularly because the data needed encompasses diverse aspects such as architectural designs, construction processes, and real-world environmental conditions. 

When those spatial models arrive, they will unlock something bigger. The construction industry currently operates in what Alain describes as a pre-mechanised mode, relying on human labour at scale. A future where robotics, offsite manufacturing and AI-driven orchestration converge could fundamentally change how the built environment is produced. 

Alain's optimism is grounded in what Buro Happold is doing internally, but the broader industry picture is more sobering. A 2025 RICS survey of over 2,200 professionals found that 45% of AEC organisations reported no AI implementation at all, and fewer than 1% had AI embedded across multiple processes. A separate Bluebeam survey of 1,000 AEC professionals put active AI usage at just 27%, with over half of respondents still relying on paper during the design phase. 

McKinsey's 2024 research underscores the scale of what is at stake: global construction output is projected to grow from $13 trillion to $22 trillion by 2040, yet productivity growth averaged just 0.4% annually between 2000 and 2022. The payoff for firms that do commit to transformation is real, but the industry has a long history of technology initiatives that fail to scale. As Bluebeam CEO Usman Shuja noted in the survey's findings, the biggest barriers in 2026 are not cost but complexity, culture and connection. 

This gap between potential and practice arguably strengthens Alain's central argument. The industry does not need another call to overhaul its systems. It needs technology that works within the mess.  

If LLMs can genuinely deliver value from existing workflows, the adoption curve could look very different from what previous digitization waves achieved. 

Alain opened the conversation by noting that the first nine weeks of 2026 felt like three years. For AEC leaders, the strategic question has shifted from whether AI will reshape the sector to whether their firms are cultivating the culture, skills, and knowledge systems to be on the right side of that reshaping. 

The Tower of Babel metaphor has haunted this industry for a generation. Perhaps the resolution was never a common language. Perhaps it was a translator good enough to make the many languages work together. 

Where does your firm sit on this spectrum? Are you building your own Google Translate, or still waiting for Esperanto? We'd love to hear your perspective. Hit reply, drop a comment, or subscribe to Bricks and Bytes to keep this conversation going. 

Watch our conversation with Alain Waha here 👇👇👇