The ConTech Exec

Insurance compliance is construction's most hated paperwork — and one of its most expensive blind spots. Vince Giacomini left a ~$50M Silicon Valley exit to go fix it.

After building a connectivity-as-a-service platform to ~$50M ARR and selling to private equity, Vince went and asked contractors what they actually hated. Two answers came back: insurance compliance and the skilled-labor shortage.

In this episode he breaks down why the GC, the sub, and the broker never share a system of record — and how his platform Forsured closes that loop and monitors policies before they lapse mid-project. Then he gets into his second bet: a construction-specific talent platform aimed at pulling the next generation into the trades.

A rare look at how an outsider investor reads construction's oldest problems.

Guest: Vince Giacomini, Founder & CEO, Forsured (former CEO, Premier Wireless Solutions)

Everyone says construction resists technology.

Olivier Grinda thinks that's a cop-out — and he sold his last company to come prove it.

A repeat founder who ran the turnaround at Clearco, Olivier crossed into construction through the Stackpoint venture studio to attack a $175B+ annual problem: rework. In this episode he argues that tech rollouts don't die because field crews are stubborn — they die because the tool solves 80% of the job and sends you back to Excel for the rest.

His rule: solve the discrete problem 100%, lean into how people already work, and never tell a 30-year superintendent how to do his job. When his team gets it right, document review drops from four days to fifteen minutes. A sharp, contrarian take on AI adoption on the jobsite.

Guest: Olivier Grinda, Founder & CEO, [construction AI startup — confirm name/spelling] (former COO, Clearco)

Construction contract risk has a direction: downhill, toward the subcontractors and suppliers least equipped to absorb it. This week Robert talks with Kendall Hoyd, founder of FairBuild AI and a 30-year operator across the trades, about the clauses that quietly take companies down. They walk through real horror stories — a $300K final payment and retention vanishing into an indemnity claim, a flooring sub forced to keep installing on a job that had stopped paying — and the false economy of skimping on the drawings budget. Kendall built FairBuild to put plain-English contract analysis in reach of subs who can't keep a lawyer on staff. It's a conversation about parity, leverage, and where the real risk actually lives.

Three takeaways:

• Pay-if-paid and indemnity clauses can be company-ending — and they usually go unread.
• Saving 25% on the drawings budget can cost multiples downstream in rework and change orders.
• The risk sits with whoever doesn't have a lawyer. Change that, and the dynamic changes.

Kendall Hoyd is the founder of FairBuild AI, a contract-intelligence platform built for trade subcontractors and suppliers.

Follow the ConTech Exec Podcast and never miss an episode.

Puneet Raj, CEO of Fieldwire (a Hilti company), joins Robert to unpack Field Intelligence — the AI platform Fieldwire launched at ENR FutureTech — and why AI in construction only pays off when it's actionable for the person standing on the jobsite at 6 a.m. They dig into a survey of 170 companies showing 72% already use AI, almost entirely for emails and meeting notes, and what crews actually want instead. Puneet argues AI won't change who makes the call — it just arms the field with the data to push back. The conversation gets into thornier ground too: who owns construction data when a job closes out, why the same data carries different valid perspectives for owners, GCs, and subs, and why the clean as-built handoff remains unsolved. Plus, the unlikely story of how Puneet went from Unilever and Amazon to digging a foundation by hand in Bosnia.

Three takeaways:

• Generic AI gets ~60–65% of the way on a construction task; the last mile is construction- and company-specific.
• Capturing senior field workers' "method to the madness" isn't about replacing them — it's about giving the next generation the why.
• "There is only one owner of the data, and that's the customer."

Puneet Raj is the CEO of Fieldwire, the field-first construction management platform now part of the Hilti Group.

Follow ConTech Exec for a new conversation every week. If this one made you think, share it with someone who runs the field.

"I Love It When A Plan Comes Together"

Hannibal Smith had it half right. A plan is a beautiful thing when it comes together — but on a real jobsite, it only comes together when the people doing the work are the ones who build it.

This episode, Robert sits down with Nialli's Nancy Knowlton and Alan, the team behind Nialli Visual Planner — digital pull planning built to mirror the Last Planner System.

Nancy co-invented the interactive whiteboard before carrying that collaboration pedigree into construction;

Alan came up through the boots-on-the-jobsite research that shaped the product. Their case is simple: the plan isn't a document handed down from an office.

It's a set of promises the trades make to each other.

We get into why so much of the industry still forces work through critical-path schedules nobody in the field ever agreed to, and what changes when you put the foremen at the wall instead.

We talk about the heroics trap — the superintendent pulling it out of the fire at the last minute — and why that runs out of road in a market with no labor to spare.

And we dig into the part most teams ignore: the data a pull plan throws off, and what it's quietly telling you about who's reliable, which region is winning, and where the next delay is coming from.

What we cover:

• How Nialli went from the inventors of the interactive whiteboard to digital pull planning
• The Last Planner framework in plain terms — should, can, will, did
• Why low-bid, transactional contracts work against the plan, and Glenn Ballard's relational-contract counterpoint
• Getting construction off heroics and onto a process that holds
• Reasons-for-variance data, regional analytics, and bidding on capacity instead of price
• New jobsite-photo and organizational-learning features — and where AI actually fits ("chat with your data," not buzzwords)

Check out Nialli at nialli.com. Link in the comments.

New episodes of The ConTech Exec Podcast every week — subscribe so the next plan comes together in your feed.

Jeevan started OpenSpace about eight years ago — his second company, built with co-founders who came out of spatial AI and computer vision. The premise is plain: no amount of reports and documents tells you what's actually happening on a job site, but a picture does. Host Robert Zimmerman sat down with him at ENR FutureTech to get into why visual data might be the truest thing in construction.

Robert's most viral article was called "Percent Complete Is Your Biggest Lie," and this conversation is why it struck a nerve. Stop trying to measure everything with money — spending $50,000 doesn't mean you got $50,000 of work done. A photo tells you what's really behind the wall, what's going up next, and where you actually are. OpenSpace pulls that visual data off 360 cameras, drones, and phones, locates it automatically, and turns it into real productivity numbers you can't argue with. The reporting most builders rely on has a garbage-in, garbage-out problem — it's the last thing anyone wants to do at the end of the day, and half of what gets written down is myth.

Jeevan's read on why good tools fail: the real cost isn't money, it's time and attention, and that's the scarcest thing in an industry where 70% of the market is mid-to-small contractors stretched thin with no R&D budget. The tools that take off don't make a fancy pitch — they take something painful and make it a hundred times easier. They get into location being everything (the superintendent doing "judging by the shadow in this photo" Sherlock Holmes work to place a defect, when a timestamp would've settled it in five seconds), and the OpenSpace origin story where the team made themselves bad interns on real projects just to learn the work.

And the part Robert's been pounding the gavel on for years: cash flow. The average sub waits 56 days to get paid, partly because nobody's fully sure the work got done — a Russian doll of trades, GCs, owners, and LPs all hedging. Staple the field reality to the invoice, take the fear out of it, and net terms can move toward zero. Cash flow is king. If everybody trusts they'll get paid, the whole job runs smoother.

Most construction robotics has lived in the vertical world — painting robots, layout machines. Rachael Ferrera of Zachry is after something else: autonomy on the heavy civil side, the yellow iron. As a self-performed heavy civil contractor with quarries, cement, and ready mix in Texas, Zachry started where autonomy makes the most sense — the predictable, closed sites — and is now pushing it onto the construction side too. Host Robert Zimmerman sat down with her at ENR FutureTech to get into what's actually working.

The thesis is simple: don't automate the operator, automate the busy work. Zachry's partnered with Bedrock Robotics on an excavator that can handle the dull, repeatable half of the job — loading dirt in a truck — so the skilled operator gets out of the cab and runs the work that takes a human: moving pipe, setting retaining wall panels, covering multiple crews. A dump run is no fun. Let the machine do that.

What sold her on Bedrock wasn't the Waymo pedigree on the team — it was that they actually showed up. From day one they wanted input from the field, not just executives: operators, engineers, the equipment team, safety, ops. Rachel's advice to any ConTech founder is the same — put your boots on, spend a day in the heat in the PPE, and learn what the work really is before you build for it. And contractors have to meet them halfway: on innovation, the high tide raises all boats, and if you hold your cards too tight the good tools get swallowed up before they ever mature.

The rest of the conversation: why this is the moment robotics finally arrives — AI, computer vision, and autonomous-vehicle hardware all maturing at once — why it's about capacity and capturing the institutional knowledge of a 30-year operator, not replacing anyone, and why adoption, not the tech, is where most of it dies. Build the best thing in the world; if it isn't easy and safe and predictable, it's useless.

Oh — and they went from first phone call to hardware on the equipment in about a week.

Steve Jones brings order to data for a living. For fifteen years, Dodge Construction Network has run a panel of 7,000 real design and construction professionals — people they know are legit and know what they're building — and gone out to take the industry's pulse on everything from change management to AI. Host Robert Zimmerman sat down with him at ENR FutureTech to get past the anecdotes and ask what the numbers actually say.

Steve's whole case is that construction is so fragmented we make huge decisions off two or three stories and call it the truth. It's usually somewhat true — and somewhat true is exactly how you burn yourself. He gets a call a week from a founder with fresh money and a shiny thing that's going to reinvent the industry, and he stops them with one question: how does this help a superintendent on a Wednesday morning at 9:15? If you can't answer that, you've got tech searching for a problem.

The rest of the conversation: why there's no silver bullet, only Kaizen — a thousand things done a little better, maybe ten thousand — and why every exhibitor promising to solve everything is setting itself up to fail. Why technology is GPS: it won't tell you where to go, it'll only get you there once you know. And the AI study he was about to present — fewer than half of contractors even know the AI features already sitting in their tools, adoption is thin on the project-management side, but for the ones who've done the real behavioral work to embed it, it's night and day, and it's showing up in their margin.

It comes back to people, the way it always does in this business. As Steve puts it: everybody wants to be first to be third. Let the other guy try it and bollocks it up — just be the fastest third.

Alan Espinoza of Reconstructive AI handed his crew ten brand-new iPads — BIM 360 loaded, models loaded, Bluebeam sessions ready. A few days later he found his superintendent in the golf cart playing Candy Crush. That's the state of construction tech adoption in one image, and it's where this conversation starts.

Host Robert Zimmerman sits down with Alan to take apart the stat everybody quotes — that 95% of tech rollouts fail — and why it's almost never the software. It's the implementation: wrong person in charge, no executive sponsorship, and nobody ever asked why the team didn't want to use the thing. Alan's premise is simple: you can't fix what you don't know is broken, so map your broken workflows first.

They get into what actually works — one-on-one over top-down mandates, showing people what's in it for them (the Universal Studios story), capturing the institutional brain before your clash-detection wizard retires off a $400M job, and the incentives that move a job site the way Duolingo streaks and seatbelt beeps move everyone else. And the one that always lands with a sub: use the tool, get tracked faster, get paid faster. Cash flow is king.

Even when the subject is artificial intelligence, it comes back to people. That's the whole game.

Todd Zabelle is back. Last time we talked flow. The time before that, implementation. This one, we put AI on the table and watched it fail at the thing everyone keeps promising it will fix.

The planning stage isn't broken. It's missing. Engineers got pulled out of the how. Tradespeople got pulled out of the what. The CM sits in the middle making bars. Three groups, three languages, nobody speaking the same one. AI doesn't fix that. AI walks in, asks for a price, and gets handed a schedule made by someone who never built anything.

Todd makes the case that AI is a subordinate tool. Useful for equipment data, dossier work, requirements management, traffic routing. Not useful for designing the production system. That still takes a human who's been on the tools and knows operations science. One plus one doesn't equal two. Capacity utilization isn't linear. The sum of cycle times isn't the cycle time. None of that is in the model.

Then we got into the data center build. Switchgear delays. Labor shortfalls. A town that kicked out its own council to stop a data center. GPU supply tightening. Banks getting jumpy on syndication. Todd's read: it's not 1999.com. It's 2012 fracking. Everybody at the feeding trough. Three years from a thinning.

The Empire State Building got built in less than a year with sketches. We can't build a tilt-up on time with 3D models and an MRP system. Todd has a theory about why. So do I.

This one's for the executives who keep buying AI as the answer without asking what the question was.