Stop Talking to Customers—Work Beside Them | Buildots

Aviv Leibovici didn't survey construction managers—he walked the floors with them and brought pastries. That on-the-ground obsession revealed chaos that no customer interview could capture, and eventually led Buildots from negative margins to tens of millions in ARR.

Key Facts

• Time spent on-site before building: ~5 months of bi-weekly site visits before shipping product (Aviv Leibovici)
• Initial product adoption failure: 90% of early users got zero value from a technically working product (Aviv Leibovici)
• First pricing miss: Quoted £4,000/month and customer signed instantly—'obviously it means we underpriced it' (Aviv Leibovici)
• Growth inflection post-pivot: From 1.5–2X annual growth to 3–4X after repositioning around proactive risk alerts (Aviv Leibovici)
• Total funding raised: ~$150 million raised; revenue in the tens of millions ARR (Aviv Leibovici)

Why Talking to Customers Is Not Enough

Customer interviews reveal how people wish their world worked—not how it actually works. Aviv Leibovici spent five months on construction sites watching managers operate before he understood what problem was worth solving. Physical presence surfaces reality; interviews surface idealization.

Before Buildots had a product, a team, or even a clear idea, Aviv was on construction sites in London carrying equipment for site managers and bringing pastries. He struck a deal with one manager: 'I walk around with him, I ask any question I want, I carry anything he needs carried, and I bring a pastry in the morning.' That arrangement gave him a front-row seat to an industry running on hearsay.

What he saw in those weekly visits couldn't have come from a survey. Managers debated whether work on the fourth floor was done or not—mid-meeting—by sending the most junior person in the room to physically go check. Project directors negotiated completion dates they knew were fiction, playing a long game of paper trails rather than managing actual progress. One manager explained it bluntly: 'All I'm trying to do is have three or four meeting summaries in which he did not deliver the thing we agreed, so I can go to his director and say you have to resource this job more.'

This is the insight Aviv carried forward into every product decision: you cannot understand the problems you can solve unless you're physically present. The construction industry didn't need better interviews—it needed someone willing to show up.

"Everyone in product says that you have to speak to your customers and I say that's bullshit. Speaking to your customers doesn't help. Because when you speak to people, you're hearing their idealization of their world. What they want it to be. When you're there with them, you see what reality is." — Aviv Leibovici

The 90% Problem: A Working Product Nobody Used

Buildots built a technically accurate AI platform that tracked construction progress in real time. But 90% of customers got no value from it—because giving people data and expecting them to change behavior are two entirely different things.

By September 2019, Buildots had a working product. Site workers would walk floors with a 360-degree camera; computer vision would analyze footage and generate a color-coded progress table—red for behind schedule, green for on track, blue for ahead (which, as early customers pointed out, never showed up because nothing in construction runs ahead of schedule).

The product worked. The adoption didn't. Aviv describes the split clearly: 'There's a difference between a product that is providing the right tool that can drive tremendous value and a product that most users will actually get tremendous value from.' Ten percent of customers used the data brilliantly. The other ninety did almost nothing with it.

The reason was behavior change. To get value, project managers had to prepare before meetings, learn a new tool, and restructure how they ran weekly stand-ups. They were too busy. Aviv realized the product's success depended entirely on him being in the room—he called himself 'the data monkey.' Once he left a project, usage collapsed. Tracking weekly minutes in the platform was measuring the wrong thing entirely. The real question was: what can we deliver in ten minutes that creates 80% of the value?

"The thing that you need is to succeed, to deliver to them what they need in ten minutes of use every week. Because that's the time they have." — Aviv Leibovici

• Computer vision tracked real-time progress across all floors and activities.
• Color-coded dashboards showed schedule status—but required users to pre-analyze data.
• Value only materialized when Aviv was physically present to guide meetings.
• Weekly usage time turned out to be a misleading success metric.

The Pivot That Unlocked 3–4X Growth

Buildots stopped selling a data platform and started selling a risk-alert system. Instead of showing project managers what had happened, it showed them what would happen if they didn't act now. That reframe—from reactive to proactive—made the value obvious and the sale easy.

The shift didn't start in a product meeting—it started with one person on Aviv's team building insights manually in Excel. Amir, now VP of Industry Transformation, pulled data from Buildots' system and built a simple model: here's your current painting rate, here's when you'll finish at this pace, here's what that means for your project. He showed it to customers. It worked immediately.

Aviv describes the concept: 'Think about something like, I tell you, you need to paint this entire building. You have one million square feet to paint. You're painting a certain amount every week. And if you continue this way, this is when you'll finish. And if you finish then, what will that do to your entire project?' Simple. Obvious. Actionable.

Before this change, Buildots was selling to construction companies running two percent margins—asking them to spend meaningful money on a tool that would help them do existing tasks better. After the pivot, the pitch became: here is a risk that will cost you far more than this product, and you can eliminate it now. Sales cycles shortened. Project directors started saying 'this is so simple' within minutes of a demo. Growth jumped from 1.5–2X annually to 3–4X.

"Suddenly, we shifted and we said: this is the performance in which you are doing, and if you continue this performance, you will finish this much late. You have a spark here, you can step on it now, or it can become a fire." — Aviv Leibovici

"Every time I speak to a project director, nine times out of ten, within minutes, they get it." — Aviv Leibovici

Prove Value in Excel Before You Build Anything

Aviv's standing rule for his product team: never send a feature to development until a customer has experienced its value without the actual product. Build it in Excel, fake it manually, or deliver it yourself. If the value doesn't show up in a hack, it won't show up in code.

This principle emerged from hard experience. The risk-alert insight that transformed Buildots' growth wasn't discovered in a product sprint—it was discovered by a team member manually building the analysis from raw data and handing it to a customer. Only after they confirmed it was valuable did the team invest in productizing it.

Aviv now applies this as a rule across the company: 'You will not send anything to development until you will show a customer that experienced this value without the actual product. Hack it in Excel, in Power BI, be the person for them instead of the product. I don't care what you do, but show that the value happens, and then go and build the feature.'

This approach also protects against the most common early-stage trap: building a technically impressive product that solves the wrong problem. The Excel-first rule forces the team to stay anchored to real customer outcomes—not engineering elegance.

"I now tell my product people: you will not send anything to development until you will show a customer that experienced this value without the actual product. Hack it in Excel, in Power BI, be the person for them instead of the product." — Aviv Leibovici

Pricing When You Have No Idea What to Charge

Buildots' first pricing was based on a rough benchmark against a manager's salary and what customers currently spent on tech. Their first real expansion deal was priced at £4,000/month—and the customer signed without negotiation. The lesson: when a customer doesn't flinch, you've underpriced.

Early pricing at Buildots was, by Aviv's own admission, mostly guesswork. The framework was loose: charge more than customers currently pay for tech, but not so much more that it creates friction. The goal was to avoid free pilots—'nobody will try to even do it if it's free'—while keeping the number reasonable for an industry with notoriously thin margins.

The reality check came fast. Aviv was pitching an expansion within their first paid project. He had a plan: open at £4,000/month extra, expect pushback, settle at £3,000. 'I'm already apologizing before I give him the number. And then I tell him £4,000. And he goes, okay, cool. So where do I sign?' He called his co-founder Roy immediately. Both agreed the conclusion was obvious: they had underpriced by a significant margin.

The lesson Aviv draws is practical: when you have no comparable benchmarks, use a customer's reaction as your data. Immediate acceptance without negotiation is a signal, not a celebration.

"I sit there and already, because I'm not a salesperson and I suck at this, I'm already apologizing before I give him the number. And then I tell him £4,000. And he goes, okay, cool. So where do I sign? I call Roy and we both agree that obviously it means we underpriced it." — Aviv Leibovici

Construction Project Management Tools: Where Buildots Fits