Is Construction Ready for AI?

Dan Laboe
Jun 12
6 min read

This is Part 1 of a two-part series on AI in construction. Part 2, The Nymbl Take: Construction's Late-Mover Advantage, dives into our thoughts on investment opportunities in AI for construction.

The Promise of AI in Construction:

AI is poised to reshape construction by addressing two critical issues: labor shortages and long-awaited operational efficiency improvements.

What is Needed:

The foundation of AI is built on high-quality, diverse, structured datasets, which are unavailable at scale in the construction space, but the AEC (architecture, engineering, and construction) space is in the process of enhancing its data culture for the future of construction.

Leveraging real-time information sources (e.g. IoT, sensors, computer vision, BIM, progress tracking solutions, etc.), employing AI-ready ERP systems to unsilo internal data (tech stack consolidation), and collaborate with other stakeholders to obtain the full scope of structure project data, will open the door for AI to drive real value in this space.

Stages of AI and Where Construction Fits:

AI development is still in its infancy, with limited memory models like ChatGPT and other generative AI solutions only recently coming to market. These solutions gave AEC leaders an understanding of basic AI models' potential. Once the foundational AI data practices are established, effective models for predictive analytics, generative AI, and other efficiency-enhancing AI functionality can be implemented.

Ideal Outcome:

The construction industry has been driven by rule-of-thumb decisions that could only be made by an AEC professional with years of experience.

Theory of Mind (ToM) AI technology, which remains a distant technological goal, will offer this industry a project copilot that can understand and anticipate human decision making at each stage of construction to work alongside industry professionals or autonomously to significantly reduce unexpected expenses, project delays, reworks, etc.

Addressing the Labor Shortage in Construction

Technology in the construction industry primarily aims to enhance productivity in existing operations, with the sector's leaders increasingly recognizing AI as a transformative solution.

According to AGC's 2025 Construction Hiring & Business Outlook report, labor shortages in construction are quickly turning from a concern to a serious problem, with 80% of AEC companies reporting difficulty in finding qualified workers to hire, and 88% predicting that it will remain difficult or even harder in 2025 (1). It’s estimated that an additional 439,000 construction workers are needed to satisfy 2025 labor demands in the US construction alone (2), a figure that is only expected to climb in the coming decade.

The AEC industry is now looking to AI to catalyze a long-awaited productivity boost and ease the strain on its understaffed workforce.

AI Investments in ConTech

($ in Millions)

Within the built environment, Construction Tech (ConTech) was the only Nymbl category (defined below) that experienced investment growth in Q1 2025, and AI-enabled solutions accounted for nearly 50% of investments in ConTech in Q1. This represents a notable 22% YoY and 18% QoQ boost in AI ConTech investments.

Nymbl's Category Definitions

Building Tech – If a building were taken off the ground, everything within that structure, including the management, maintenance, operations, and even the materials it's made out of, would fall into this category. Building tech encompasses solutions for developers, owners, operators, underwriters, and brokers of commercial, industrial, and residential buildings. (e.g., HVAC systems, electrical components, structural components, energy management, renovations, facility management, tenant management, VPP, commercial real estate, development, digital twin, carbon management, new building materials, etc.)

Infrastructure Tech – All technologies related to the maintenance, management, and optimization of horizontal assets (roads, bridges, utilities, power, water, etc.) as well as supply chain management and advanced manufacturing. (e.g., utility management, grid management, road & bridge inspection, water/waste management, clean energy generation, EV charging, smart city technologies, etc.).

Construction Tech (ConTech) – All technologies involved in the construction of vertical and horizontal assets, and removed at the end of a project. (e.g., project management, field management, preconstruction solutions, workforce management, industrialized construction, insurtech, financing, architectural & engineering solutions, etc.)

Many advanced manufactured technologies could be utilized in one, two, or all three of the above categories, but from an investor perspective, Nymbl classifies specific startup technologies into one of these three buckets based on its primary application.

Corporate venture arms in construction are becoming hyper-focused on AI-enabled startups, backing 85% of the Q1 investments into this niche – up dramatically from just 33% in Q4 2024.

Corporate-Backed AI Investments in ConTech

($ in Millions)

It is encouraging that corporate investors are beginning to take a dominant role in the development of the AEC’s AI startup ecosystem. Channel partnerships are critical to an AI's success, and obtaining the required data to build AI tools is challenging for startups in construction.

AI in construction is in its early stages of development and will require the industry to change its data culture. Corporate collaboration—particularly through the sharing of proprietary data—will be critical to unlocking the full potential of AI in the industry.

The barrier to forming these mission-critical partnerships is significantly reduced when corporations have a vested interest in the startup’s success and the value of its resulting solution.

AI Data Efficacy - The Foundation of Future AI Tools

The “5Vs of AI Data Efficacy” listed below are the foundational elements upon which future AI tools can be built (both for construction and elsewhere):

5Vs of AI Data Efficacy

1. Volume – the amount of data

2. Variety – the types of data being collected

3. Velocity – speed at which data is generated

4. Veracity – accuracy and reliability of data

5. Value – the usefulness and relevance of the data to operations

The good news is that the built environment generates a vast quantity of diverse and complex data, and we believe a cultural shift is happening that augurs well for the future of AI in the industry.

Corporate investors poured hundreds of millions into data-generating and data-enhancing startups in Q1'25, as their applications begin to yield a quantifiable competitive edge. This shift is underscored by the growing level of early-stage investments into technologies like IoT sensors, reality capture, BIM, document management, ERP systems, and other data collection/management solutions, which are making strong data practices second nature across the industry.

Corporate-backed Investments in Data-Enhancing Startups

($ in Millions)

Strategic investments into data management startups reached an all-time high in Q1'25 and were nearly double the total investments in all of 2024.

These real-time data collection technologies introduce a new layer of structured project data with the granularity required for emerging AI to deliver real economic value.

The commitment from construction corporations to enhancing their data collection and management practices is the first critical step to building an AI-optimized industry.

Lots of Legacy Data - No Easy Way for AI to Access

As noted above, a key concern around AI technology in construction is its dependence on large volumes of high-quality data. This raises a critical question: Is the construction sector set up to support complex AI/ML models?

Virtual design & construction (VDC) platforms, such as building information models (BIM), have become widespread in the AEC space, with more than 80% of contractors leveraging some form of digital construction technology (3). These digestible digital databases are gold for the future of AI in the AEC space.

However, these databases remain largely siloed within AEC corporations that treat these datasets as invaluable proprietary digital assets. While this is a justifiable position, these firms risk slowing the development of transformative AI models by limiting access to their data.

AEC leaders need trusted channels to share data anonymously, and certain founders of applied-AI solutions are creating collaborative data structures in their backend systems that leverage proprietary data confidentially. Convincing executive leadership in large corporations to participate in data collaboration partnerships will be a long-term battle.

Simulated Data – AI Founders’ Cheat Code?

As real-world data remains scarce or difficult to standardize in the fragmented AEC landscape, many players developing AI solutions are turning to proprietary simulation software as a workaround to fill the critical need for AI-ready inputs.

Simulations of real-world conditions and scenarios can generate an unlimited amount of data. However, building an effective simulation requires a sizable amount of initial data and perfect specifications, making it time-intensive and expensive endeavor.

The "Theory of Mind" Hurdle in ConTech AI

The entire construction process is driven by a culmination of human-driven ‘judgment calls’ that require years of dynamic professional experience. These judgment calls are driven by “rules-of-thumb” reasoning —instinctive, experience-based choices that are critical to today’s building process.

These nuanced, situational decisions remain difficult for a machine to make today, and the degree of reasoning required for AI solutions to replace human professionals (Strong AI) remains in very early development. AI specialists are uncertain of the timeline for achieving “Theory of Mind” in AI technology (aka true artificial intelligence), with predictions for technology maturation ranging from 5 to 25 years.

Conclusion

Startups should prioritize building flexible data integrations that can scale across disparate data types and structures while seeking partnerships that grant them access to richer datasets. On the other side of the table, corporations would benefit from shifting their view on data as a proprietary asset to a strategic enabler and invest in systems that standardize, structure, and ultimately share value-creating data anonymously across the value chain.

AI in construction will require a collaborative data culture, where value is created not by hoarding information but by making it actionable at scale. Nymbl's thoughts on investable opportunities in AI can be found in Part 2 of this article: The Nymbl Take: Construction's Late-Mover Advantage.