Building Smarter
How AI and Data Could Unlock a New Era for Architecture
Written by Andrew Carton, WG’26
I encourage you to take a moment and think about your high school experience – the classes you took, the teachers who made an impression, the lessons you learned and those you wish you had. At the public high school I attended, we had the standard fare: math, science, history, English. But there were also artistic requirements and real-world preparatory classes. Despite my inability to carry a tune, I sang in the school chorus. I reluctantly entered the art fair with an abstract portrait of a friend’s sibling. And I took a class called Life and Technology, which gave us hands on exposure to practical skills that might serve us later in life.
In that class, I thought deeply about architecture for the first time. We competed to design and build wooden truss bridges, judged on both strength and style. We used AutoCAD, a 3D computer-aided design (CAD) tool, to draft and refine our concepts. The software allowed us to see our ideas take shape and test different variables before committing to tangible form. My bridge looked good, failed fast, and taught me something more enduring: design lives at the intersection of creativity and physical constraints.
Though I abandoned any dreams of becoming an engineer, I developed a quiet obsession with the built environment. I started browsing the websites of large architecture firms, scrolling through renderings of stadiums, transit hubs, and office towers. If you haven’t done this before, I recommend it. Look up Populous, HOK, or HKS and explore one of their upcoming projects. These digital mockups are rich with possibility – proposals for what a space could become and how it might shape those who experience it. What captivated me then still resonates today: the sense that design can give form to imagination and redefine how we interact with the world around us.
More than a decade later, I’ve returned to this interest from a new vantage point. Architecture, like law – another deep-seated interest of mine – is a high-stakes, data-rich profession. Both are shaped by precedent, creativity, and client expectations, but remain structurally siloed and resistant to new technology. Architecture is poised for transformation, but progress faces an uphill battle against misaligned incentives, fragmented data infrastructure, and a dependence on legacy tools.
As I’ve explored technological change across professional services, I’ve developed a deep appreciation for vertical-specific innovation – products tailored to the needs of a particular field. I appreciate the natural logic these solutions require: industries have specific challenges, and people who create products to address those challenges must understand those needs intimately. This fall, I dove into the world of Legal AI and spoke with generous founders, academics, and practitioners. The law is an industry known for resisting change, and yet, meaningful pockets of innovation have taken hold. I found the same cautious optimism in architecture, a similarly conservative and risk-averse profession. This post outlines five key lessons exploring how AI is starting to reshape the field of architecture.
Autodesk: The Legacy Giant
To understand the current state of architecture tech, start with Autodesk. The company launched AutoCAD in 1982, offering a PC-based tool to digitize manual drafting. By the early 2000s, AutoCAD had become the most widely used design software in the world, with users ranging from architects and engineers to city planners and interior designers.
Autodesk doubled down on its influence with the 2002 acquisition of Revit, a company developing building information modeling (BIM) software. Unlike traditional CAD tools, BIM allowed architects to create data-rich models that could be shared with other stakeholders in construction and engineering. CAD facilitated 2D or 3D drawings based on lines, while BIM enabled architects to visualize buildings through smart, interactive objects. Adoption of BIM has accelerated since, with Revit emerging as the dominant standard: one recent survey found that 45% of European architects use Revit as their primary BIM tool.
In 2020, Autodesk acquired Spacemaker, a Norwegian startup focused on AI-powered generative design. Spacemaker was eventually folded into Autodesk Forma in 2023, a cloud-based platform designed to help teams collaborate on early-stage design decisions. Forma promises to evaluate site conditions, enable more seamless project coordination and recommend potential reconfigurations; whether it delivers on that promise, however, remains up for debate.
Autodesk remains the architecture industry’s default operating system – deeply entrenched, widely used, and increasingly difficult to dislodge. Its position has been secured by decades of product expansion, smart M&A, and high switching costs experienced by users operating in a deeply interconnected ecosystem. Autodesk’s market cap hovers around $50 billion, and the company spent nearly $1.5 billion on research and development in its latest fiscal year. Yet for all this investment, the pace of meaningful innovation has slowed.
Competitive Landscape: New Tools, New Entrants
Despite its scale and influence, Autodesk is under pressure. In 2024, activist investor Starboard Value disclosed a $500 million stake in the company and accused it of misleading investors during an internal accounting probe. In its March 2025 letter to shareholders, Starboard argued “Autodesk’s long history of financial and operational underperformance has led to a severe lack of management credibility and significant shareholder frustration.” Autodesk has since laid off 9% of its workforce and is preparing for a potential proxy battle.
Having seen similar shareholder battles firsthand, I know how disruptive they can be. In theory, public companies can handle governance fights while staying focused on innovation – Starboard’s engagement has yet to make a material dent in the company’s share price. In practice, however, attention is diverted, risk tolerance falls, and teams slow down. Autodesk has a reputation for tinkering with its go-to-market strategy, and Starboard’s involvement might risk exacerbating those concerns for customers.
More critically though, while Autodesk Forma is meant to represent the future, many architects say the tool still feels hypothetical. Revit, meanwhile, was supposed to be a shared data backbone for the entire design-to-build lifecycle. Instead, it functions more like next-gen drafting software – a powerful upgrade to AutoCAD, but not the transformative platform it aspired to be. This disconnect has become increasingly obvious to firms. As one architect told me, “We have 800 BIM models in one space. You need conformity. But we’re held hostage by the current technology.”
Against this backdrop, startups are rushing in. A tracker labeled “AI in AEC” – operated by industry observer Stjepan Mikulic – now lists over 1,500 tools focused on Architecture, Engineering, and Construction. Among the most notable is Motif, founded by Autodesk’s former co-CEO and product CTO. They share deep domain expertise and trust with enterprise clients, two factors needed to sell products in architecture. Motif has raised $46 million from CapitalG and Redpoint, offering a browser-agnostic collaboration layer that streams directly from Revit and supports real-time sketching and feedback.
Motif is part of a broader movement toward cloud-native, collaborative, and interoperable design tools. These features are not just buzzwords – they reflect deep frustrations with Autodesk’s legacy approach. Revit, for all its capabilities, historically relied on file-based work sharing and still suffers from cumbersome data exchange. Many newer tools aim to fix this.
Still, the holy grail is interoperability: the ability for systems and stakeholders to seamlessly share data and context across tools, teams and phases of a project. Revit promised this but never fully delivered. What we’re seeing now is a renewed push to fulfill that vision.
Interoperability: Elusive But Critical
The challenge isn’t just technical – it’s structural. Architecture operates within a complex project delivery model made up of three main actors: clients, architects, and constructors. These parties have distinct responsibilities, incentives, and risk profiles. Architects generate ideas, builders execute them, and clients evaluate them on price and vision. Architects bristle against constructors who override ideas; builders dismiss architects who ignore material constraints.
Business models reinforce this tension. Architects are often paid a percentage of construction costs, and clients typically select architects based on low first cost (i.e. lowest bid). This disincentivizes upfront transparency: detailed models and robust data sharing may reduce errors and save money later, but they also lower construction costs – and therefore, architect fees. Builders, in turn, profit from change orders and unexpected conditions. Detailed models and clear channels for data sharing eat into that margin, reinforcing a feedback loop that stymies interoperability. As a result, nobody has a strong financial reason to share clean, coordinated data from the outset. Even though full interoperability would reduce errors, shorten timelines, and unlock new value, current incentives keep data siloed.
Organizations like buildingSMART are trying to solve this through open standards like IFCs (Industry Foundation Classes), which aim to create a universal data language. Construction tech giant Procore likens IFCs to PDFs: “It’s a neutral file schema used across the AEC industry […] It’s a static version of a model that can be moved from one platform to another.” Startups like Speckle are taking a more developer-centric approach to interoperability. The company builds data hubs that support multi-tool collaboration and real-time data flows. Speckle raised $12.5 million in October from Addition, Foundamental and others to expand its platform.
Yet most of these tools still serve the architect more than the broader project delivery ecosystem. They improve design workflows but don’t address the deeper disconnect between design and construction. To unlock real transformation, we need systems that help each party understand the other’s constraints, preferences, and priorities.
Data: A Foundation Yet to be Poured
Currently, architecture lacks a centralized data hub. Whereas the legal field has a robust corpus of data in the form of Westlaw or LexisNexis, knowledge in architecture is poorly organized and virtually inaccessible. Data exists in heterogenous forms – drawings, spreadsheets, etc. – with levels of abstraction and complexity. To create effective technological solutions for architecture, akin to applications gaining traction in other professional services, data first must be centralized and prepared for training purposes.
This is where AI can help. AI systems are uniquely good at extracting structure from unstructured information. They can index, sort, and cross-reference fragmented project data to build a shared understanding across teams. Done right, this could bridge the gap between design intent and construction execution – surfacing insights that reduce rework, improve buildability, and preserve the architect’s original vision.
Phil Bernstein, an architecture technologist and Deputy Dean at Yale’s School of Architecture, has proposed a non-profit data trust to unlock this potential. The idea is to anonymize and pool data from architecture and construction firms, under the supervision of an independent third party. Once cleaned and centralized, firms that contribute to the data trust receive access to data driven insights, and the data can help train next-gen tools. If successful, it would mirror the role of Westlaw or LexisNexis and catalyze step change innovation rather than incremental improvement on existing software.
While I support the proposal, I also recognize its practical difficulty. An independent data trust will struggle to overcome the intractable business models that disincentivize data sharing. Data aggregation in the law provides an alternative to an independent data trust. Westlaw and LexisNexis are neither independent nor nonprofit: they are subscription services owned by Thomson Reuters and RELX, respectively.
I suggest there’s a simpler solution than a data trust. Rather than fight Autodesk’s dominance, startups should lean into it. Autodesk could launch a shared data platform where contributors earn credits or discounts for submitting anonymized models. This platform would offer benchmarking tools in exchange, delivering immediate ROI by helping firms bid more effectively and reduce costly errors. Over time, it could become the default data layer for the AEC ecosystem, unlocking a new generation of intelligent applications at the intersection of architecture and construction.
Innovation Today: Possible Without Interoperability
Even without a shared data layer, there’s plenty of opportunity for startups and incumbents to make architecture smarter. Tools like qbiq – which raised $16 million from Insight Partners in January – radically simplify the early design phase. The company allows users to generate floorplans and 3D visualizations for commercial real estate, from simple prompts like seat count or finish material. An Israeli architecture firm, Gindi Studio, used qbiq to tackle a surge in client requests for layouts and space planning. By integrating the tool into day-to-day operations, the studio increased its preliminary planning capacity by 300%, allowing the team to deliver test fits without pulling resources away from ongoing projects.
This illustrates the near-term commercial value of AI tools that simplify and accelerate decision making while enabling architects to tackle more – or higher value – work. Other architecture firms are working directly with cloud platforms like AWS and Azure to develop domain-specific AI features. Emerging tools tend to address larger TAMs such as commercial or multi-residential real estate. For companies operating in spaces like sports and entertainment, where there are fewer projects and domain-specific tools, direct partnerships make strategic sense.
Architecture firms are also starting to monetize their predictive capabilities. Just as investment banks charge higher fees when sales or IPOs outperform expectations, architects could charge based on how accurate their energy or usage predictions prove over time. Architects constantly
make predictions about the outcomes of their work: with AI to improve modeling and benchmarking, they can better monitor and monetize the accuracy of those predictions. They can strengthen their financial position by linking fees to impact. This approach sidesteps the misaligned business incentives that have historically discouraged architects from embracing new technology. By monetizing their predictive insights, architects can evolve their role from creative service providers to strategic forecasters, redefining how they create and capture value.
Architects also have room to deepen project engagement after handoff. By using IoT data and user feedback, firms can act as partners to owners and operators – fine-tuning layouts, optimizing performance, and shaping real-time decisions that affect the bottom line. They can power business intelligence dashboards that inform decisions ranging from layout adjustments to refresh strategies in venues like arenas. This turns architects from static contributors to dynamic
collaborators, aligning them more closely with building management software and driving better outcomes for clients throughout an asset’s lifecycle.
Finally, AI can automate core design tasks. While today’s tools help catch code violations or clashes, tomorrow’s will flag operational inefficiencies, recommend layout changes, and guide compliance in real time. Even without a shared data layer or interoperability, these systems can draw on vast datasets of similar projects within a firm’s archives. With firm-specific data, AI applications can evolve from identification to active design recommendation. And these systems won't replace architects – they'll make them faster, more precise, and more focused on the high value work humans are best suited to accomplish.
Conclusion
The architecture industry sits at a crossroads – still anchored by legacy tools like Revit, yet increasingly aware of the need for more modern, flexible and intelligent solutions. Many professionals feel constrained by the limitations of existing platforms and the lack of seamless interoperability. Architecture remains deeply collaborative, but misaligned incentives between architects and builders have led to siloed data and inefficient workflows. Without a central data infrastructure, the industry struggles to fully leverage AI, automation, and other innovations transforming adjacent fields like legal services.
Despite these structural challenges, there are clear signposts for progress. Cloud-native, collaborative, and interoperable tools improve how teams design and communicate, while firms actively experiment with AI to enhance the accuracy and marketability of predictions. Untapped potential exists in helping architects extend across the building lifecycle, using data to bring design beyond ideation.
Ultimately, a massive opportunity lies in building shared data ecosystems that unlock smarter design and delivery for all. Until then, plenty of near-term innovation can elevate the profession and bring new value to clients. Solutions require coordination across stakeholders, and likely a dose of creative pragmatism. But the momentum is real. The next chapter of architecture won’t just be visualized. It will be structured, optimized, and intelligently built.
If you’re building in the space or want to discuss anything written, I’d love to hear from you!