Enterprise

Home Remodeler West Shore Runs on AI Despite Hallucinations

A billion-dollar bathroom company deploys Claude and computer vision across 3,200 employees—but keeps humans in the loop when the model confuses millimeters with dollars.

Omega Editorial· July 5, 2026· 4 min read

A trades business that went AI-native

West Shore Home, a central Pennsylvania remodeling company, has quietly become one of the most aggressive AI adopters in the construction industry. The firm generated $933 million in gross revenue in 2025 and now operates at a $1.15 billion annual run rate, completing bathroom and kitchen renovations with a tech stack that includes computer vision, LIDAR scanning, and large language models embedded in nearly every workflow.

Founder B.J. Werzyn, a Penn State engineering graduate who abandoned aerospace for home improvement in 2006, has built a 3,200-employee operation that treats remodeling like a systems engineering problem. His company uses a proprietary iPad application called Hawkeye to generate 3D scans of customer bathrooms in under a minute. Those scans feed structured data—precise measurements, fixture locations, obstacles—directly into an internal configure-price-quote system called Felix, eliminating the transcription errors that plague traditional contractors.

In 2026, roughly 70% of West Shore's in-home sales appointments generated a Hawkeye scan, rising to three-quarters for bathroom-specific visits. Only 1,600 customers out of 136,000 appointments declined the technology. The company employs 115 people on proprietary systems, including 23 focused exclusively on AI.

When Claude thinks EBITDA is measured in millimeters

Werzyn's enthusiasm for AI is matched by his candor about its limitations. He described a recent exchange with Anthropic's Claude model while building five-year growth forecasts. After a detailed conversation about adjusted EBITDA projections, the model suddenly declared that fiscal year 2028 would deliver "260 millimeters of EBITDA."

"It clearly knows those are dollars," Werzyn said, surprised the model could grasp context yet revert to the wrong unit.

That kind of hallucination has shaped West Shore's deployment strategy. The company developed conversational SMS agents to follow up on sales leads, achieving higher response rates than phone calls and generating about 10% of all appointments. But Werzyn has resisted scaling the system more aggressively. If 10% to 15% of customers experience an off interaction—a misscheduled appointment, an inaccurate quote—the reputational cost in a high-ticket category is too steep.

As a result, AI runs the numbers, populates forms, surfaces options, and suggests installation dates. But humans still visit homes, verify scans, confirm quotes, and hit "schedule." The technology is everywhere; it just doesn't close the loop alone.

Scaling revenue without proportional headcount

West Shore added roughly 600 net new jobs over the past three years, bringing total employment to more than 3,200, including 1,209 installers and 657 design consultants. Werzyn believes AI will let the company reach $2 billion in revenue with around 6,000 employees instead of the 7,000 it would have needed without automation. The last-mile work—installation, plumbing, physical labor—remains human. The overhead supporting it gets streamlined.

The company received a 20% investment from private equity firm Leonard Green & Partners in 2020, with Werzyn retaining 80%. He also joined the board of Utz Brands in 2024, bringing his operational lens to another legacy business navigating technology and consumer shifts. West Shore has donated to Penn State's Beaver Stadium renovation and secured field naming rights, while recruiting computer science and AI talent from the university.

Why it matters

West Shore's approach illustrates how AI is being absorbed into capital-intensive, labor-heavy industries far removed from Silicon Valley. The company isn't replacing workers with models—it's using models to reduce coordination overhead and error rates in a $500 billion sector still dominated by clipboards. But Werzyn's insistence on keeping humans in the loop, even as AI handles more tasks, reflects a pragmatic calculus: in high-trust, high-ticket categories, the cost of a hallucination isn't just a bad output—it's a lost customer.

These details were first reported by Fortune.

#ai deployment#construction technology#anthropic claude#computer vision#home remodeling#ai hallucinations

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Enterprise

Enterprise· 3 min read

NHS to Deploy AI Triage Tool on Patient App Across England

£10bn digital overhaul includes AI-powered appointment routing and consultation recording, but health leaders urge caution on privacy and productivity claims.

Via AI Watch · Jul 5, 2026
Enterprise· 3 min read

Why AI Adoption Fails: The 52-Point Trust Gap Between Leaders and Workers

Employees use generative AI three times more than executives realize, yet only 1% of companies have integrated it into daily work—here's what's broken.

Via AI Watch · Jul 5, 2026
Enterprise· 3 min read

SK Telecom Plans 15GW AI Data Center in South Korea by 2035

The telecom giant aims to position Korea as an Asian AI infrastructure hub amid global computing shortages.

Via AI Watch · Jul 5, 2026