The ConTech Exec

Construction AI is moving from hype to real operational leverage, and Raymond Zhao thinks the biggest unlock is not replacing engineers, but giving them a smarter assistant that catches problems before they become RFIs, change orders, and field delays.

Raymond, co-founder of Structured AI, joins Robert to explain how agentic AI reviews drawings, detects cross-disciplinary inconsistencies, and flags code issues inside complex Revit and BIM workflows—without taking over professional liability.

They discuss why Raymond left architecture, how Structured AI is improving construction quality control, why AI adoption depends on change management, and what it means to treat a building set like an intelligent asset you can actually ask questions of.

This episode is essential listening for anyone interested in architecture, engineering, construction, AI adoption, and the future of infrastructure.

is the co-founder of

, a construction AI company helping engineers and builders detect issues faster, reduce review burden, and improve drawing QA.

Construction fabrication is still riddled with a costly lie: the model is not the workflow.

David Basiji explains why the real bottleneck is getting from BIM and CAD into the shop and then through every cut, bend, weld, and handoff without losing time, accuracy, or visibility.

Robert and David trace the unlikely path that led David from bioengineering and the Human Genome Project into construction tech, then into PipeServer - a platform built to connect design tools, fabrication equipment, and the field. Along the way, they expose why “paper to PDF” is not digitization, why exact lengths can be the wrong answer in fabrication, and why the smallest workflow mistake can ripple into massive delays.

Discover:

• Why fab shops need a live system, not static cut lists and spreadsheets
• How PipeServer moves data from Revit, AutoCAD, Tekla Structures, and SDS2 into machines and shop workflows
• Why field fit, bend stretch, and fabrication sequence details have to be handled intelligently
• How cloud-connected tracking prevents down-rev files, duplicate cuts, and security headaches
• Why AI is not the right tool for deterministic shop workflows, but is powerful for analyzing bottlenecks and productivity data

David also breaks down how PipeServer expands beyond a single pipe cutter brand to support the entire fab shop, from manual tools to automated machines, and why the biggest opportunities may be with mid-market and smaller shops that have been underserved by legacy software.

If you care about prefab, BIM, production efficiency, or the future of construction data flow, this conversation is essential listening.

David Basiji is at PypeServer, where he helps build software that connects design, fabrication, and field execution. He brings deep technical experience from bioengineering, machine development, and building scalable products.

Perfect for contractors, fab shop leaders, VDC teams, and construction tech buyers who are tired of process chaos and want a clearer, faster path from model to field.

We close out the three-part series with Joe and Ryan from Cavtera, and this is the one that actually answers the question the first two episodes set up. Turns out the code you build inside Palantir Foundry belongs to you — Joe compares it to Microsoft Word, nobody thinks Word owns your essay. So who's really steering a company's tech? Not IT, not finance — the guys actually paving Main Street. We get into why Thomas Cavanagh Construction Limited named their whole operating system after Tom himself (there's a story here involving a Volkswagen, a sandwich, and a highway shoulder that's worth the price of admission alone), and why a schedule that only lives in a binder never poured a yard of concrete. I leave you with the real question: are you willing to define your own business, honestly? That's the price of admission before any of this works.

Part two with the Cavtera guys, Joe Patrois and Ryan Quesnel. Last week we cleared up the vocabulary — what Palantir Foundry actually is, what an ontology is (the nouns and verbs of your business, nothing fancier than that). This week we get into the bruises.

Joe and Ryan tell the origin story: two project management guys at Thomas Cavanagh Construction sitting in a boardroom, griping, until one of them said "nobody's coming to save us." So they saved themselves. Spreadsheets first. SharePoint and Power BI next. And every new tool they bolted on fixed one thing and loosened something else — like fixing a leak on iron pipe. Tighten this joint, spring a leak two fittings down. That's tech debt, and most contractors are drowning in it without ever calling it by name.

Then Palantir showed up at their shop for a boot camp, and they spent the next stretch of their careers building what became TOM — Total Operations Management — the operating system that now runs Cavanagh and became the product Cavtera sells today.

We also get into the stuff that actually kills software rollouts: forgotten passwords, phone screens zoomed to twelve thousand percent, and the difference between tracking productivity and tracking progress (hint: they are not the same thing, and your budget is neither one). Plus my pitch on why you should stop counting grains of sand and start tracking your downtime instead.

If you missed Part 1, go back and watch it first. Part 3 drops next week.

Real People, Real Work, Real Talk.

The conversation delves into the concept of Palantir Foundry and its impact on construction technology. It explores the challenges of legacy technology, change management, and the role of AI in modernizing construction operations. The conversation delves into the challenges and opportunities in the construction industry, focusing on the adoption of technology, the impact of AI, and the need for seamless integration of digital tools with physical operations. The discussion emphasizes the importance of understanding the true ontology of the construction business and the need for a holistic approach to technology adoption.

Takeaways

• Palantir Foundry is an ecosystem that allows businesses to organize and assimilate all their data, enabling the creation of custom applications and solutions.
• Change management is a critical aspect of implementing new technology, and the process involves untangling legacy systems and processes to build tools around the best way to work. Technology adoption in the construction industry presents challenges and opportunities.
• Understanding the true ontology of the construction business is crucial for successful technology integration.

Chapters

• 00:00 Tackling Change Management and Solutions Engineering
• 45:22 Understanding Ontology and System Elements

Before he was Field COO for public sector at Procore, AJ Waters was an engineer whose actual job on a project was moving RFIs out of his company's spreadsheet and into the owner's spreadsheet, then carrying the answer back. His words: "I was the API between the two systems."

He says that's still the reality on almost every job site.

This week we get into public sector construction, where every dollar spent on software is a dollar a taxpayer thinks should've gone to a project. Owners maturing off spreadsheets. Data centers as infrastructure, and the Nebraska power plant that had no customer until one showed up. And AJ's bigger argument — that construction spent twenty-five years digitizing paper and called it transformation.

Plus the Empire State Building, built in under a year with everybody working out of the same building. Which last time I checked is not how we do it anymore.

New episode every week.

Steve Jones started studying BIM in 2005, back when five people would give you ten definitions of it. Twenty years later he's still doing the same job: finding out what contractors, owners, and designers are actually doing, and what's actually working.

His verdict on two decades of construction tech promises is not what you'd expect. The vendors weren't the problem. The industry doesn't want to be reinvented — it wants to be better in about 10,000 specific ways. What got underestimated was how hard organizational change is when every project is a temporary organization stitched together from firms with different contracts, incentives, and digital maturity.

We cover the Procore study on expertise, the change order research showing only 7% of GCs rate themselves very effective at managing T&M work, why 38% of projects run over schedule while only low-20s run over cost, and what separates owners who think tech means software from owners who think tech means better information.

Steve Jones is Senior Director, Industry Insights Research at Dodge Construction Network. Free research at construction.com.

#ConstructionTech #ConstructionManagement #ProjectControls #ConstructionIndustry #Preconstruction

Everybody quotes the failure stat. Almost nobody reads it right.

Alan Espinoza of Reconstructive AI returns for the full Texas Drive Time conversation, and he opens by reframing the number the industry hides behind. It isn't that seven out of ten implementations failed. It's that seven out of ten people struggled to use what you bought them. Different problem. No software fix.

Then the harder question: what are we actually training these models on? An architect, a GC, and an owner will each get a different answer out of the same tool, because each one is prompting it to defend their own position. If AI learns the industry as it behaves today, all we've built is a faster way to point fingers.

Host Robert Zimmerman and Alan cover the administrative state built out of CYA contract language, the 96% of data that goes straight to the shelf, CPM schedules as a backwards pass from a date corporate already picked, and what walks out the door when your one indispensable guy retires.

Plus a Reconstructive AI update — Y Combinator, GC incubators, and agents crawling county permit records to find work early.

🎧 Also on YouTube.

Real People. Real Work. Real Talk.

#ConTech #AIinConstruction #ConstructionTechnology #ConstructionProductivity

Your construction schedule is detailed to the hour. Your procurement schedule is two lines on a Gantt chart. That's the problem.

Jeff Kaufman, founder of JIT Pro, started as a field carpenter, ran large projects as a superintendent, and moved into management where he traced nearly every late job back to the same root cause: procurement. In this episode, Jeff explains why expediting procurement doesn't save time — it compounds risk that lands three months later when the steel package shows up wrong. He and Robert get into hidden buffers versus strategic buffers owned by the whole project team, the silent transfer of risk that happens when contractors sign unrealistic schedules, and why working backward from required on-site dates often proves the buying should have started before the contract was even signed. They also cover the Living Building Challenge's double procurement process, cognitive overload on project teams, and who JIT Pro is built for: growth-stage GCs jumping from two jobs a year to five.

Guest: Jeff Kaufman, Founder, JIT Pro (jit-pro.com)

#ConstructionTechnology #Procurement #ConstructionPlanning #GeneralContractor #ConTech

Materials are roughly half the cost of any construction project, and on most jobs they get less management attention than the coffee order.

Eshan Jayamanne, Founder & CEO of Krane, joins Robert to talk about why the construction supply chain stayed broken for so long — and what changes when jobsite delivery data starts flowing upstream into preconstruction.

They cover the whiteboard that started it all, why hospitals are quietly fighting data centers for the same transformers and switchgear, and why the right time to find your bottleneck is before you pour, not after.

Eshan also explains why Krane refuses to pitch itself as "AI for anything" — lead with the problem, and let the tool be the how.

Three takeaways:

• Your procurement log is a diagnostic. Columns, macros, submittal ties — the spreadsheet tells you exactly how a team actually works.
• Design around the supply chain. Same cost, two supplier paths, six months of schedule difference. That decision belongs in precon, not month nine.
• The tools now train the people. GCs are using these systems to onboard new project engineers faster than the jobsite learning curve ever allowed.

Eshan Jayamanne is a licensed PE and third-generation builder who worked at Stantec, WebCore, Boldt, and Chevron before founding Krane, now deployed across major healthcare and data center programs.

Listen, follow the ConTech Exec Podcast, and share it with someone still running their supply chain off a whiteboard.