VERTICALS is sponsored by Parafin.
The leader in embedded finance for vertical.
Click here to learn more.
Scott Wolfe built a category-defining construction payments business in Levelset. He was just hitting his stride in 2021 when Procore, the 800-pound gorilla in construction SaaS, showed up out of the blue. He went from first call to $500M exit in months.
On this episode of VERTICALS, we unpack how to navigate a big exit as a founder. How should you think about relationship-building with big potential acquirers—if at all? How do you respond to the first M&A offer? Where do you draw the line between “I know we can keep growing and want to keep running this thing” and “this would be life-changing for my employees, shareholders, and me?”
In a lot of ways, Scott is an unconventional founder. Before Levelset, he was a lawyer by trade—though never really in spirit or ambition. Born and raised in New Orleans, he founded his startup about as far outside the “coastal VC” circuit as possible. His growth story was uneven by all measures: as we dissected with Levelset CRO Martin Roth in a prior episode, $0-1M and $10-20M took no time at all, but $1-10M was a long, many-year journey. He didn’t have a longstanding relationship with Procore, who ended up acquiring them in a flash over a few months in late 2021. Scott wasn’t looking to sell. We wasn’t the best networked. 99% of the time, he was just heads down scaling a valuable business.
Was Levelset in the right place at the right time? Did Scott’s intuition about the importance of payments in a big industry—before the space got “hot” in software—pay off? Did he navigate the interplay with systems of record expertly? Or is this story just another great reminder of why “great companies are bought, not sold?” All are valid takes… you be the judge.
Beyond M&A, we also go deep on construction as a vertical: where the real white space is (hint: Scott doesn’t think it’s another job-site tool for GCs); what distinguishes a point solutions and an extensible wedge product; and, of course, how AI is changing the market for technology in the construction, trades, and blue-collar industries. We also hit a Vertical Pulse covering raises from Trunk Tools, AIM, and PermitFlow + our Vertical Playbook for the week, inspired by LevelSet: Identifying a “lightning bolt” vertical workflow.
Happy listening + join us on YouTube for much more!
I) Vertical Titan
Scott Wolfe — Founder & former CEO @ Levelset
The backstory
Scott’s path into construction tech was anything but linear: entrepreneurial family, early web tinkering, law school, then Hurricane Katrina turning New Orleans into an unusually intense construction ecosystem.
His “spark” was brutally simple: filing a lien — a small, common, high-stakes action — was weirdly difficult, even for professionals. He started building tooling around that gap, and over time it expanded across states, use cases, and customer types.
The early years were slow and scrappy. Scott ran it “on the side of his desk” for years, got it to roughly $200K in revenue, then went all-in around 2012.
The hardest part
Turning an initially transactional “wizard” product into a real SaaS model (and learning what customers would pay for repeatedly)
Explaining a niche, legally-heavy workflow to outsiders (investors, recruits, and even the broader market)
Surviving on a small early round ($500K) for multiple years while bouncing in and out of cash-flow positivity
Breaking the industry’s gravitational pull toward the big general contractors and staying committed to a bottoms-up thesis
The $1-10M ARR stretch: that took multiples of $0-1M ARR and $10-20M ARR
Memorable lines from Scott
“Lien rights run through [construction] like a lightning bolt — every project, of every size, of every shape.”
“We got beaten down for years on market size… [too many] people are building companies for the VC, not for the customer.”
On the decision to sell: “It gets hard to say ‘let’s just keep going’ when you see what it means for people who’ve been with you for eight or nine years.”
The wedge: liens as a “lightning bolt” through every project
Scott framed lien rights as a workflow that cuts across project sizes, stakeholders, and geographies — a rare trait in construction.
That’s what made Levelset legible as more than a feature. It wasn’t “paperwork automation.” It was payment certainty in an industry defined by payment anxiety.
Go-to-market: bottom-up, not GC-first
While much of construction tech targets general contractors, Levelset went bottoms-up: subcontractors and material suppliers as the beachhead.
A key nuance: “bottom of the chain” doesn’t mean “small.” Levelset had thousands of SMB customers — but also enterprise supplier contracts that reached seven figures. The category was different; the buyer profile could be very big.
Capital: raise enough, then execute
Scott is openly skeptical of raising massive rounds “because cycles come and go.”
His path was the opposite of hype:
years of compounding and product learning
a small angel round that lasted an almost comical amount of time
then venture rounds once the model was tighter and the team understood the problem deeply
How the Procore deal happened (fast)
Scott had met Procore years earlier around a potential partnership when they explored a payments product and didn’t land it.
Then in July 2021, they reappeared — and quickly made the conversation explicit: they tried building it, it didn’t work, and now they were hungry to buy. Scott wasn’t for sale, but the combination of deal momentum, stakeholder impact, and strategic vision made it hard not to engage.
Once it started, it moved like an “asteroid”:
July: first serious acquisition conversation
By September: under terms
By Thanksgiving: closed
Founder psychology: the real stress isn’t diligence
Scott’s surprising takeaway: the hard part wasn’t the back-and-forth of terms.
The hard part was deciding whether to end the journey. He was growing faster than ever and enjoyed being CEO. He had a bigger vision he wanted to keep pursuing. But when he saw on paper what the outcome would mean for employees who’d been loyal for nearly a decade, he realized ego was holding him up. Procore’s initial offer he rejected outright. But they were super aggressive, and when that second offer came in much stronger, he realized ego might’ve been playing too big a role.
He also noted the macro context: in his view, the market environment in 2021 wasn’t going to last forever. That influenced the calculus. He admits he had no illusion of being able to time the market, but it turned out that way. Market cycles matter enormously. As he shared, you need to be both good and lucky—just one won’t work.
Where Scott sees white space in Construction AI
Scott’s argument is straightforward: the industry is far bigger than commercial GCs, and many adjacent stakeholders don’t even live in the “construction tech” universe.
Areas he’s excited about:
material suppliers (huge, under-networked, often unaware of the mainstream construction stack)
restoration and non-commercial construction workflows
the professional layer around construction (architects, engineers, insurance, bonding)
cross-cutting workflows like permits that naturally touch multiple parties
And some parting founder advice:
Scott continues to feel that way too many founders build for what VC currently care about, rather than for their customer. That’s inherently illogical because what investors care about changes by the year, and ultimately—if you build a big, high-growth business—VCs will start to care. Especially in the era of Vertical AI, TAM opportunity is changing faster than TAM perception. Your job (as a founder) is to know the problem and the customer better than anyone else. If you can get conviction that there’s a big business to be built starting in a place that others think is too small... that’s a good thing.
II) Vertical Market Pulse
Trunk Tools: Is “ChatGPT for the jobsite” defensible?
Trunk Tools raises $40M | Kyt Dotson @ SiliconANGLE
Trunk Tools is a clean example of Vertical AI done with ambition: connect the messy universe of construction docs and systems, then ship a conversational interface that a field team can actually use.
But the strategic question is the same one we debated on-air: if your core value is “connect your documents and chat,” how do you defend against horizontal platforms that want to sell the exact same promise to every enterprise?
Founder implication: if you’re going “platform-first,” you still need a few hair-on-fire workflows that become your anchors — the things customers can’t live without, even when budgets tighten.
AIM: Physical AI hits the built world
AIM raises $50M | Rachyl Jones @ Semafor
AIM is tackling an old, legacy corner of the built world: heavy machinery that hasn’t changed much in decades. Their bet is that autonomy doesn’t need brand-new fleets — it can be retrofitted into what’s already on-site.
We like this as a “physical AI” wedge, but it comes with obvious scaling constraints (hardware, deployment, support). The upside is equally obvious: if labor shortages and safety issues keep compounding, the value per deployment can be enormous.
Founder implication: hardware makes everything harder — so make sure the pain you’re solving is permanent, not cyclical.
PermitFlow: Permits as an “authoring layer”
PermitFlow Raises $54 Million Series B | Francis Thumpasery @ PermitFlow
Permitting touches everyone — contractors, developers, architects, and municipalities. That surface area is the feature, not the bug.
Scott’s take (and we agree): workflows that run through every project — in every geography — are where platforms can actually emerge. Liens did that for Levelset. Permits can do it for PermitFlow.
Founder implication: instead of starting with “the user” (often the GC), start with the workflow that forces multiple stakeholders to show up. That’s where network effects and expansion paths get real.
III) Vertical Playbook
Identifying a “lightning bolt” workflow
Why it works
In vertical software and AI, the biggest opportunities often hide inside workflows that—while they seem too niche to outsiders—are non-negotiable to insiders.
Levelset is focused on liens, which have a powerful property: as Scott put it, they run right through industry workflows “like a lightning bolt, every project of every size of every shape.” That creates repeatability, urgency, and a natural expansion path across stakeholder types. It allowed Levelset to address every ACV and GTM motion in the book over time. While this seems table-stakes, there’s an important distinction
Instead of conceptualizing his approach as “let’s sell software to GCs,” Scott and team thought about it this way: “let’s own a workflow that is not only critical to GCs but also a through-line to the entire vertical ecosystem” (subs, suppliers, lenders, and courts are all already forced to participate).
That’s why:
Bottom-up sales worked from the start, liens being inextricably tied to how contractors get paid (eventually, all ACVs were relevant)
Network effects emerged without being a marketplace
Procore couldn’t just “build it” (they tried, failed, then paid up to acquire the best)
How to run it (Levelset-inspired blueprint)
Target a workflow that the entire ecosystem must participate in (often but not always financially related)
On wedge, start with the smallest painful action that’s common and time-sensitive (the thing people hate doing, but can’t ignore)
Ship a narrow product that wins trust fast, even if it begins transactionally
Use real usage to discover what should become subscription vs. “one-off”
Expand state-by-state / jurisdiction-by-jurisdiction as a moat (complexity is your friend, creating a moat naturally)
Start where pain is greatest but grow up- / down-market without changing the wedge: SMB volume + enterprise accounts can coexist if the workflow is universal.
Recognize there may be a “growing pains” period in which you have initial PMF but the network effects haven’t kicked in yet; often this will be an issue of perfecting your sales funnel and understanding which ACVs work best early
Founder litmus tests
Does your workflow unite vertical stakeholders who don’t otherwise naturally coordinate around a single source of truth?
How (if at all) do stakeholders communicate and collaborate today?
Do stakeholders use different approaches to solving problems around this workflow today (usually this is a good thing)?
Why would these secondary stakeholders also be willing to pay (or do they already) for solutions around this workflow?
Can you own the workflow for your primary stakeholder from day one?
What does it take for you to say confidently: if we disappeared tomorrow, it would make life harder for the entire ecosystem?
Does the involvement of other stakeholders set you up for potential product and TAM expansion down the road?
Is there jurisdictional or other complexity that will contribute to the moat around your source of truth and / or data asset over time?
What we debated on-air
Whether “platform-first” Vertical AI interfaces will hold up as horizontals get better at knowledge + chat
Why construction founders keep defaulting to GC-first go-to-market (and why that may be a dead end)
The founder trade-off in M&A: maximizing value vs. continuing the mission — and how stakeholder outcomes change that equation
Next week on Verticals
Tune in next week for a monster of an episode covering Vertical AI Systems of Record with Nate Baker, founder / CEO of Qualia and co-founder of Fractal.
