The Back Office Becomes the Battleground

Trend Spotlight

The construction industry's AI story keeps shifting. Two years ago, the conversation was all about drones mapping sites and robots laying bricks. A year ago, it was generative design and BIM automation. This week, two funding rounds tell a different story — capital is flowing into the operational plumbing of construction. Not the glamorous end. The messy, compliance-heavy, insurance-laden, payroll-nightmare end.

Trayd raised $10 million in three weeks to automate payroll, HR, compliance, and labour cost tracking for specialty trade contractors — the electricians, plumbers, ironworkers, and concrete crews who do the actual physical building work [1]. These businesses have largely been invisible to the technology boom. They're running on spreadsheets and phone calls, navigating union rules and multistate labour law without any of the digital infrastructure that general contractors have been building for years. Y Combinator, which had already backed Trayd, doubled down. That signal matters.

Meanwhile, Shepherd raised $42 million — a Series B — to solve an even more obscure but increasingly urgent problem: insuring AI infrastructure construction [2]. Every GPU cluster needs a building. Every data centre project needs insurance before it can break ground. The current process is weeks of calls, emails, and manual analysis by carriers using disconnected tools. As Shepherd's chief executive put it, "The AI race has moved from the cloud to the construction site." The whole AI buildout depends on an insurance market that's been operating the same way for decades. Shepherd is betting AI can fix that.

These aren't flashy bets. They're bets on the infrastructure of the infrastructure — and increasingly, that's where the money is going.

This Week's Headlines

🟢 Trayd Raises $10M Series A for Construction Back-Office Automation

Trayd, a New York-based startup building a back-office operating system for specialty trade contractors, raised $10 million in a Series A round led by White Star Capital, as reported by Crunchbase News. The round closed in three weeks and included participation from Y Combinator (doubling down on a previous bet) and Suffolk Technologies, along with new strategic backer RXR Realty [1]. That brings Trayd's total funding to $17 million since its 2021 founding.

The company targets the specialty trade segment — contractors who place skilled workers on sites to do the actual physical work. Its product automates payroll, HR compliance, and labour cost tracking, addressing the pain of managing union rules, multistate labour laws, and back-office processes that eat into margins already measured in razor-thin percentages.

Why it matters: Specialty trade contractors have been an underserved market in construction technology. General contractors have benefited from multiple rounds of digital transformation; the trades that build for them have largely been left behind. Trayd's quick raise and strong backer list suggests this is a real, validated pain point.

🟢 Shepherd Raises $42M to Insure AI's Physical Infrastructure

Artificial intelligence-native insurance provider Shepherd raised $42 million in a Series B round led by Intact Private Capital, with participation from Spark Capital and Costanoa Ventures, SiliconANGLE reported. The round brings total funding to $67 million. Shepherd underwrites commercial construction insurance — specifically targeting the wave of data centres, power plants, and chip fabrication facilities being built to support the AI boom [2].

The problem Shepherd is addressing is genuinely underappreciated: AI infrastructure projects cannot break ground without insurance in place. The current process involves contractors shopping for quotes across multiple carriers, each using disconnected tools and manual analysis, and it routinely takes weeks. Shepherd says it uses AI to accelerate underwriting, and states it has grown revenue by more than seven times over the past two years, off an undisclosed base.

Why it matters: This story sits at the intersection of AI infrastructure investment and construction. As hyperscalers, chipmakers, and AI labs pour hundreds of billions into physical infrastructure, the ability to insure those projects quickly becomes a genuine constraint on how fast the industry can build.

🟡 AI Converts Paper Drawings to BIM at 98.8% Precision

Researchers at Suzhou University of Science and Technology published a study in February 2026 presenting DBAL-YOLO — a deep learning framework that takes scanned 2D engineering drawings, including pre-digital drawings from the 1980s, and automatically generates complete 3D BIM models. The system reported 98.8% detection precision and 98.3% recall, trained and tested on a dataset of 3,960 annotated drawings, according to an archBIM.cloud analysis of the peer-reviewed paper, published in the Scopus-indexed Smart Construction journal [3]. It works independently of commercial BIM software, generating 3D solids directly from 2D pixel data.

The technical approach combines three stages: YOLO-based structural element detection enhanced with Dynamic Snake Convolution (for narrow, elongated shapes like beams), OCR-based geometric correction that reads dimensional text and aligns elements to architectural modules, and a Python-based 3D reconstruction engine.

Why it matters: Vast archives of existing building stock — hospitals, government facilities, legacy infrastructure — exist only as paper drawings. The ability to digitise these at scale without manual modelling effort has significant implications for retrofit planning, facilities management, regulatory compliance work, and digital twin creation for older building stock. This is research from a credible academic institution, published in a peer-reviewed journal, and worth tracking as it moves from lab to production.

🔶 Intuit Claims Launch of AI-Native Construction ERP

Intuit (NASDAQ: INTU) states it has launched a construction edition for its Intuit Enterprise Suite platform, describing it as "an AI-native, end-to-end ERP built specifically for the complex realities of the $2 trillion construction industry," according to a press release from Intuit's investor relations team [4]. No independent product review is yet available.

Given Intuit's scale and brand recognition through QuickBooks and Mailchimp, entry into construction ERP is strategically significant if the product delivers on its claims. However, all capabilities described — AI-native, end-to-end construction management — are self-reported by the company, and no independent evaluation or customer case study has been published.

Why it matters: QuickBooks is already used by many small contractors. An AI-native ERP from Intuit could displace dedicated construction software providers if it delivers on its stated capabilities. Worth watching for independent reviews.

🟡 Acumatica 2026 R1 Embeds AI Risk Detection Into Construction Workflows

Acumatica's 2026 R1 update to its cloud ERP adds AI-driven forecasting, anomaly detection, and progress billing automation directly into its construction edition workflows, covering the project lifecycle from bid through closeout, ERP Today reported. The release is aimed at helping contractors identify issues earlier and act with greater precision under tight margins, unpredictable material costs, and evolving compliance requirements [5].

Why it matters: This is the second construction ERP AI announcement in the same week, alongside Intuit's claim. The pattern suggests mainstream construction software is now embedding AI as a standard feature rather than a premium add-on. Mid-market contractors who haven't evaluated their ERP stack recently may find themselves falling behind peers using AI-augmented financial forecasting.

🟢 73% of AEC Professionals Still Not Using AI — But Attitudes Are Shifting

A global survey of 1,000 AEC professionals conducted by Bluebeam and reported by the American Society of Civil Engineers found that only 27% currently use AI in their operations. More strikingly, 52% still use paper during the design phase, and 43% rely on physical signatures and approvals [6]. These figures suggest the AI adoption gap isn't primarily a software problem — it's a data and process problem.

But attitudes are changing. Fifty-six percent of survey respondents believe AI will compensate for the ongoing construction skills shortage. And among the 27% already using AI, 94% plan to expand that usage in 2026.

Why it matters: The industry's relationship with AI is becoming bifurcated. Early adopters are already seeing measurable returns and expanding fast, while the majority is still working out how to get its own data infrastructure ready for AI. The competitive advantage from early adoption looks real, and likely to compound.

🟡 ConTech Startup Funding Hits $1 Billion in Q1 2025, Up 46% Year-on-Year

Construction technology startup funding reached $1 billion in the first quarter of 2025, a 46% year-on-year increase, according to data cited by USGlass Magazine — making ConTech the only built-environment category to grow both year-on-year and quarter-on-quarter in the period [7].

Why it matters: At a time when broader real estate technology investment has been retreating, sustained capital flow into ConTech signals investor confidence that construction remains under-digitised relative to its size — consistent with the operational-layer bets described above.

Data Point of the Week

56% of AEC professionals believe AI will compensate for the ongoing shortage in construction skills.

Source: Bluebeam global survey of 1,000 AEC professionals, reported by ASCE Civil Engineering Source, December 2025 [6]. 🟡 REPORTED — Bluebeam is a vendor in the AEC technology space; its survey carries commercial interest, but was reported by ASCE, an independent professional body.

Construction has faced a labour shortfall for years — skilled tradespeople are retiring faster than they're being replaced. The fact that most survey respondents see AI as the compensating mechanism, not just a productivity improvement, indicates a shift in how the industry frames the technology: not "AI instead of people" but "AI to do more with fewer available hands."

The Longer View

The AI Insurance Bottleneck

Every AI data centre, power plant, or chip fab needs insurance in place before construction starts. Traditional underwriting takes weeks of manual analysis. As AI infrastructure investment accelerates, insurance processing speed is becoming a genuine constraint on how fast the industry can build [2].

Digitising Legacy Building Stock

The DBAL-YOLO research suggests a near-term capability to digitise vast portfolios of buildings — hospitals, government facilities, ageing infrastructure — that exist only as paper drawings. The immediate applications are in retrofit planning, facilities management, and digital twin creation for older building stock [3].

ERP Vendors Converge on AI

Two construction ERP vendors announced AI-driven features in the same week. The pattern suggests embedded AI is becoming a standard expectation in construction software rather than a premium add-on — worth watching for how the larger platforms respond [4][5].

Sources

[1] Crunchbase News — "Exclusive: YC Doubles Down On Trayd, A Construction Tech Startup That Just Raised $10M In 3 Weeks," https://news.crunchbase.com/venture/construction-tech-automation-trayd-ai-seriesa/ — 2026-03-25. 🟢

[2] SiliconANGLE — "Insurance tech startup Shepherd raises $42M to underwrite the physical layer of AI," https://siliconangle.com/2026/03/24/insurance-tech-startup-shepherd-raises-42m-underwrite-physical-layer-ai/ — 2026-03-24. 🟢

[3] archBIM.cloud — "AI Creates BIM Models from Paper Drawings — 98.8% Precision, Zero Manual Modelling," https://archbim.cloud/en/blog/ai-creates-bim-from-paper-drawings-2026 — 2026-03-24. 🟡

[4] Intuit Investor Relations — "Intuit Launches New AI-Powered Construction Edition for Intuit Enterprise Suite," https://investors.intuit.com/news-events/press-releases/detail/1302/intuit-launches-new-ai-powered-construction-edition-for-intuit-enterprise-suite — 2026-03. 🔶

[5] ERP Today — "Acumatica Embeds AI Risk Detection Into Construction ERP," https://erp.today/acumatica-embeds-ai-risk-detection-into-construction-erp — 2026-03-25. 🟡

[6] ASCE Civil Engineering Source — "Architecture, engineering, construction sector slow to adopt AI, survey shows," https://www.asce.org/publications-and-news/civil-engineering-source/article/2025/12/18/architecture-engineering-construction-sector-slow-to-adapt-ai-survey-shows — 2025-12-18. 🟢

[7] USGlass Magazine — "ConTech Startup Funding Surges 46% in Q1 2025, Hits $1 Billion," https://www.usglassmag.com/contech-startup-funding-surges-46-in-q1-2025-hits-1-billion/ — 2025. 🟡