Construction Drones and Progress Tracking at OpenSpace

Every construction project generates a version of the same problem. Somewhere between what was planned, what was built, and what was documented, information gets lost. A superintendent knows what is happening on their floor. The project manager in the trailer has a different picture. The owner reviewing a PowerPoint in a conference room has a different picture still. And by the time a dispute emerges, the truth of what actually happened on the job site is already gone.

Jeevan Kalanithi has spent the better part of a decade building technology to solve this problem. As CEO and co-founder of OpenSpace, he leads one of the most widely used reality capture and construction progress tracking platforms in the industry. His Foundamental University masterclass is one of the most detailed and honest accounts available of what it actually takes to build technology for an industry that does not work behind screens.

This article draws entirely on that conversation.

Watch the full masterclass here: Product for the Field - a deep dive into construction progress tracking

From Construction Drones to Something Better

Before OpenSpace, Jeevan was president of 3D Robotics, a drone company that had pivoted from consumer hardware toward enterprise construction applications. The experience shaped everything that came after, and not in the way you might expect.

Construction drones, he found, were genuinely useful. In drones construction contexts, flying over a job site and generating a map of progress from above gave superintendents a capability they had never had before. The ability to see what was happening without being physically present, and to have visual proof of site conditions at a specific point in time, was valuable enough that customers paid for it.

But construction drones also had a fundamental limitation that no amount of engineering could fully solve.

"Drones are good for ground-up projects," a superintendent told him. "But most of what we do isn't even ground-up. And it's only useful for innings two and three of a building. Indoors is where the action is."

The second limitation was operational. On a busy job site, with superintendents and assistant superintendents working twelve-hour days, flying a drone is one more task on a list that is already too long.

"Which exact person is going to go and fly the drone on Friday?"

Jeevan recalls thinking, looking around job site trailers. "These are all very busy people. They're trying to build a building, and flying a drone is yet another thing on the list."

This led him to a principle he returns to throughout the masterclass:

"If you're going to solve someone's problem, do it in a way that removes work, doesn't add it."

Construction drones solved a real problem. But they added work in the process of solving it. That gap became the founding logic of OpenSpace.

The Insight That Built OpenSpace

The founding call came from Mike Fleischmann, one of Jeevan's co-founders from the MIT Media Lab, who was trying to find an apartment from across the country and wished he could see what spaces looked like without having to be there. Mike and their third co-founder Philip DeCamp had been experimenting with 360 cameras as a way to create immersive visual records of spaces.

Jeevan's immediate reaction was to redirect the idea.

"That's a cool idea that might benefit you personally, Mike. But the people who could really use it are enterprises, people who build buildings. For them, a picture really is worth a thousand words."

The core thesis was simple: construction is an industry where the most important information is physical and visual, but the tools available for capturing and sharing that information were either inadequate, too complex, or too burdensome for the people doing the actual work. Construction drones had pointed toward the solution but hadn't reached it. 360 cameras, combined with the right software, might.

What OpenSpace needed to do was make visual documentation of a job site as frictionless as possible, so that the capture happened as a byproduct of normal work rather than as a separate, scheduled activity. The technical challenge this imposed on the team was significant.

"Mike, Philip, and I kind of already knew that was how it had to work. We weren't happy about it because it was technically much more difficult. But in a way we were happy, because we knew we had the technical ability to pull it off better than most people, and that if we did, we'd have a real competitive advantage."

Validating the Idea: The $50,000 Story

The earliest validation of the OpenSpace concept came from a superintendent named Tim, at a time when the platform had exactly two job sites and could only be operated by the three founders themselves.

Tim had gotten into a dispute with a drywall contractor who claimed to have completed work on a specific date. Tim disagreed. The two were deadlocked over a potential $50,000 change order. Then Tim remembered that the project had OpenSpace running. He pulled up the platform, navigated to the exact location in question, and used what he called "instant replay" to review the visual record from the disputed date.

He was right. The work had not been completed when the contractor claimed it had been. The change order was avoided. The dispute was resolved in minutes.

"I asked: was the guy you were fighting with angry about this?" Jeevan recalls. "He said: no, not at all. As long as you have the facts, we all work together every day. If the facts are there, people own up to what they did and move on."

For Jeevan, this was more than a customer success story. It was proof that the core concept would scale. If a three-person company with a prototype platform could save $50,000 in ten minutes on a single job site, the value of comprehensive visual documentation across thousands of projects was enormous.

Construction Progress Tracking: The Bigger Picture

The dispute resolution use case was where OpenSpace started. Construction progress tracking is where it grew.

About a year into the company's life, Jeevan attended an OAC meeting on one of their customer projects. OAC meetings, where owners, architects, and contractors review project status and resolve open issues, are a fixture of construction project management. They are also, in Jeevan's description, frequently painful.

"A project engineer would create a PowerPoint or a spreadsheet, they'd go down a list of issues, and these meetings could be really painful. People would dispute every description. There was always this moment of tension when something was going to cost somebody money. The truth was out on the job site, but you were in the trailer, and there was this awkward question: are we all going to leave the meeting and go look at who's right? Nobody wants to do that."

Then he attended a meeting where OpenSpace was on the screen instead of the spreadsheet.

"They were just clicking around and looking at each issue with visual evidence of what was actually there. The meeting was so much faster. And there was no tension. People could just see what was going on."

This was the moment that shifted Jeevan's understanding of what OpenSpace was building. It was not a documentation tool with a dispute resolution use case. It was a new way of running construction projects, where reality data replaced description-based records as the primary medium for tracking progress, resolving issues, and making decisions.

OpenSpace's Progress Tracking product is the most direct expression of this vision. Using the visual data captured through regular site walkthroughs, the platform generates high-level construction progress tracking insights about project status: whether the schedule is on track, where delays are likely to emerge, and how productivity compares across different areas of the project. It can automatically update P6 schedules, with human verification to ensure accuracy.

"That's the insights layer for executives or regional directors," Jeevan explains. "But even in the field: imagine someone in room 603 trying to solve an issue. Spatial AI agents can flag issues relevant to that person based on their location, and even suggest who should resolve it."

Spatial AI: The Next Wave After Construction Drones

The evolution from construction drones to 360 capture to spatial AI is, in Jeevan's telling, a single continuous story about getting construction teams closer to the reality of their projects.

Construction drones gave teams a top-down view. OpenSpace's 360 capture gave them a comprehensive indoor view. Spatial AI, the next phase Jeevan describes in his masterclass, will give them an intelligent view: one that not only shows what is happening but understands it, flags what matters, and surfaces the right information to the right person at the right moment.

"All the AI developments that everybody's been reading about, generative AI, large language models, are fundamentally oriented around language, text, and documents. They aren't necessarily great at understanding real physical reality. That's not ideal for our customers, because what they do is manipulate real physical reality."

The gap between where AI currently is and what construction needs is significant. But the dataset OpenSpace has accumulated across thousands of projects, and the location-awareness infrastructure the company has built, including a technology called Auto Location that can determine the position of a smartphone indoors without GPS, positions them to close that gap in ways that generic AI platforms cannot.

"Spatial AI isn't just about what's in front of the camera at a given moment. It also has to incorporate all the prior visual information in that space, pull in language documents like specs and drawings, and understand location. As one customer told me: if I don't know where something happened, I might as well not know it at all."

The field agent use case Jeevan describes is concrete: a person standing in a specific room on a job site, with an AI agent that knows where they are, knows what the visual history of that space looks like, and can surface punch list items, outstanding issues, and relevant specifications before they even ask.

"Building these field agents that allow people in the field to have relevant information right away, acting as if they have 15 more years of experience than they actually do — that's a really big deal."

Staying Customer-Focused at 350 People

One of the most practical sections of Jeevan's masterclass concerns how OpenSpace has tried to preserve the customer proximity that defined its earliest days as the company has scaled to 350 people.

The early approach was simple and direct. Jeevan and his co-founders spent time as informal interns on job sites, sitting in on meetings, watching how people worked, and asking questions. The insight that shaped OpenSpace's core UX, that capture needed to happen passively during a normal site walk rather than as a dedicated activity, came directly from a project manager on a job site in Berkeley telling him that nobody had time to do it any other way.

"That experience in the field was critical to the founding of OpenSpace. If we had just been thinking about it in an office, we wouldn't have committed to building the technology the way it needed to be built."

At scale, the mechanisms are more systematic but the principle is unchanged. Product managers are expected to have direct relationships with customers they can call for feedback. The leadership team maintains frequent customer conversations. And OpenSpace actively hires former customers into solutions engineer roles, creating an internal feedback loop that can validate product decisions in seconds.

"If a product manager here doesn't have a handful of people in their contacts they can just call to get feedback, that's not acceptable."

Two Failure Modes in Construction Technology

Jeevan's framework for why construction technology companies fail is one of the clearest articulations of a pattern that repeats across the industry.

The first failure mode is familiar: technologists who build impressive technology that nobody in construction actually wants. They understand the technology but not the industry, and the product never finds real demand.

The second is more subtle, and more common among founders who come from construction itself. "They're almost never wrong in identifying a real problem. They felt it themselves, they're scratching their own itch. But oftentimes they create what I'd call low-ceiling companies. They solve a problem that turns out to not be a very big problem, or it's really just a feature of another system. And they can't get past $5 or $10 million in revenue."

The escape from both failure modes requires holding two things simultaneously: enough domain knowledge to understand what customers are actually saying, and enough distance to imagine solutions that go beyond what customers would think to ask for.

"Take feedback seriously, but not always literally."

Cultural Relevance: The Unexpected Advantage

One of the most memorable moments in Jeevan's masterclass is a conversation with a superintendent who had been in the industry since the 1960s. The superintendent praised OpenSpace not for its technology but for the way of working it enabled.

Before digital tools, he explained, if there was an issue on the job site, he would gather the relevant people, point at the problem, and they would fix it. Now, even minor issues require forms, descriptions, back-and-forth messages, and weeks of resolution time. OpenSpace's image-first approach felt, to him, like returning to the simpler, more direct way of working he remembered.

"What you guys are doing is this image-first way. It's like the old way. You just look at the problem, deal with it, solve it, and move on."

For Jeevan, this moment crystallized something he had not fully articulated before: OpenSpace was not just making construction more efficient. It was participating in a movement within the industry toward reclaiming a way of working that technology had inadvertently made harder.

"Maybe we're part of a simpler way to build, something that touches not just on our technology but reminds people of a recent past that was a bit more efficient, a bit less wasteful, when buildings got built more quickly."

He calls this cultural relevance, and argues that it is one of the most underrated forces in construction technology adoption. Products that tap into something the industry already values, that feel like a natural extension of how builders think about their work rather than an imposition from outside, have a fundamentally different adoption curve from products that are purely efficiency plays.

Conclusion: From Drones to Intelligence

The line from construction drones to 360 capture to spatial AI and construction progress tracking is not a series of separate trends. It is a single story about an industry learning, slowly and sometimes painfully, to make decisions based on the reality of what is actually happening on its job sites rather than on descriptions, documents, and approximations of that reality.

OpenSpace sits at the center of that story. Jeevan Kalanithi's Foundamental University masterclass is the most complete account he has given of how the company got there: the lessons from 3D Robotics, the founding insight, the early validation stories, the customer proximity that shaped the product, and the spatial AI roadmap that defines where it is going.

The full conversation is freely available at university.foundamental.com/masterclass/jeevan-kalanithi.

Related Articles: More from Foundamental University

All 13 masterclasses from Season 1 are freely available at university.foundamental.com. Other sessions relevant to construction technology and project management include:

Dimitrie Stefanescu (Speckle) on open source construction and rebuilding the data layer of AEC

Scott Wolfe (Levelset) on construction payment, risk, and empowering underserved stakeholders

Matthias Tauber (BCG) on leadership that scales inside complex construction and infrastructure organizations

Q&A

Question: Why didn’t construction drones become the definitive solution for progress tracking?

Short answer: They helped, but added work and missed where most work happens. Drones offered valuable top-down, outdoor views—useful mainly on ground-up projects and early/mid “innings” of construction. But most activity (and many disputes) occur indoors, where drones struggle. Operationally, asking already overextended superintendents to schedule and fly drones added tasks instead of removing them. OpenSpace’s guiding principle emerged here: solve problems in ways that take work off the field, not add to it.

Question: What was the key insight behind OpenSpace’s approach, and how was it validated?

Short answer: Make capture a byproduct of normal work using 360 cameras and smart software. Instead of scheduling drone flights, OpenSpace designed passive capture during routine site walks—frictionless for field teams, though technically harder to build. The concept was validated early by a superintendent who used OpenSpace’s “instant replay” to resolve a $50,000 drywall dispute in minutes, proving that comprehensive, time-stamped visual truth could unlock immediate, material value.

Question: How does OpenSpace’s Progress Tracking change OAC meetings and day-to-day project management?

Short answer: It replaces description-heavy debates with shared visual facts and actionable insights. In OAC meetings, teams can click through time-aligned imagery of the exact area in question, speeding decisions and reducing tension. Beyond meetings, Progress Tracking analyzes site walkthrough data to show schedule health, likely delays, and area-by-area productivity, and can even update P6 schedules with human verification. Executives get an “insights layer,” while field teams receive location-relevant flags and suggestions that help resolve issues faster.

Question: What does “Spatial AI” mean in this context, and how is it different from generic AI?

Short answer: It’s AI that understands the physical jobsite—what’s where, when, and what it means—rather than just text. While most recent AI advances center on language and documents, construction needs systems grounded in reality data. OpenSpace is positioned for this with a vast visual dataset, location-awareness infrastructure (including Auto Location for indoor positioning without GPS), and the ability to fuse historical imagery with specs and drawings. The vision: field agents that know where you are, understand the room’s visual history, surface punch items and relevant specs, and feel like having 15 extra years of experience in your pocket.

Question: Why does “cultural relevance” matter for construction tech, and how does OpenSpace exemplify it?

Short answer: Tools that fit how builders naturally work adopt faster than tools that fight the culture. A veteran superintendent praised OpenSpace’s image-first workflow as a return to the old, direct way: look at the problem together, fix it, move on—without layers of forms and back-and-forth. By aligning with that familiar, efficient ethos, OpenSpace isn’t just an efficiency play; it taps into values the industry already holds, accelerating trust and adoption.