Marco Herbst (00:00)
Product-market fit? It was getting an angry phone call at two o'clock in the morning from Texas because the system wasn't working. And when you compare that to the two or three years prior where our system wasn't working but nobody noticed.

Roland Siebelink — Host (00:44)
Hello, welcome, everyone. Here's yet another episode of Breakthrough AI Operators. And today I'm super honored to have a guest all the way from Europe. He co-founded one of Ireland's most visited websites in the early Internet era, and he sold it before most people had even heard of SaaS. And then he spent five years in Berlin doing nothing, by design. And when he came back, he spotted a gap that most software founders would never even look at. How do you keep a camera alive on a construction site for two years without maintenance? Not AI, not analytics, just make it work. Twelve years and 13,000 years of construction video later, that boring obsession has become a construction intelligence platform trusted by Intel, Microsoft, Meta, and Shell. And with that everyone meet my guest Marco Herbst. Welcome to the show, Marco!

Marco Herbst (01:42)
Roland, thank you very much for that introduction. Great to be here.

Roland Siebelink — Host (01:46)
Yeah, it's an honor to have you. You're dialing in from Dublin, is that right?

Marco Herbst (01:50)
I'm in Wicklow, which is just about one hour south on the east coast, one hour south of Dublin.

Roland Siebelink — Host (01:56)
Let's talk business Marco. You went boring on purpose. Hardware reliability before AI. Was that conviction or was it constraint?

Marco Herbst (02:05)
Yeah, that's part of our story. Really, it's about the people we were selling to. I've always been interested, myself, my business partner, in cameras as a communications tool and all the great things that you can get from it. But you can't get much if it doesn't work. And making it work in a construction environment is surprisingly hard. A lot of our initial energy was just making sure that we were able to deliver that part of the puzzle. And I think in hindsight, that's actually what separated us from a lot of other people who also understood the potential of the technology. In a world where there's a construction project that's got a 12-month or an 18-month deadline, nobody's interested in waiting 24 months until it works. You can say it was by constraint.

Roland Siebelink — Host (02:44)
By constraint, exactly. Well, tons of questions already arising from that, but tell us the story how you even landed in construction, because that was not the business you were traditionally in, right?

Marco Herbst (03:01)
You mentioned in the intro that I spent five years doing nothing in Berlin. I'll have you know I was very busy doing nothing. I filled every day. It just maybe wasn't all productivity orientated. From my bedroom window, I could see this gardening project, the Princess Garten that the guys are putting together, and old piano factory that was being — I started to make a time lapse video. IP cameras were just arriving then. This is around 2010. And Axis, the Swedish company, had just made their first model — these little IP cameras. Things were moving to the cloud. That was one of the big words in tech. The AI of its day was cloud-based software. And so we put those two ideas together and we were pulling images to the cloud. We'd built a really nice UI, which is very unusual in the world of cameras. And I knew we had something, but at the time what I got wrong was I didn't think it was about looking at construction sites. We went in the direction of security, which seemed more logical. It seemed like the bigger existing market. I quickly learned that that wasn't my place.

Roland Siebelink — Host (05:06)
So just kept on developing out the technology, found your way back to construction?

Marco Herbst (05:22)
Hit construction and immediately knew we had it. We were solving very real, very important problems. You had real-time visibility into the construction project. We were able to do it with high resolution video. We were able to record the video and solving real problems. By that stage I was back in Ireland and we had American clients who were doing big investments in Irish construction projects. They were able to use Evercam. They became really dependent on it and we knew we had something good and we've worked from there.

Roland Siebelink — Host (05:50)
Now you mentioned product-market-fit and knowing when you have it. Michael Seibel of Y Combinator often says one of the biggest problems is every founder thinks they have product-market-fit long before they actually have it. What's your definition of product-market-fit and how do you know if you have it?

Marco Herbst (06:13)
It was getting an angry phone call at two o'clock in the morning from Texas because the system wasn't working. And when you compare that to the two or three years prior where our system wasn't working but nobody noticed. That was really depressing. Getting the angry customer who really cared and who was really using the system and needed it back online, that's when we knew we had something and something really valuable.

Roland Siebelink — Host (06:25)
Is that a similar definition you would apply to a startup that you might invest in as an angel? Can you always apply that definition to any tech business?

Marco Herbst (06:48)
Dunno. I haven't heard it said before. But it's a reasonable test. Turn it off and see if someone cares. There have been so many essential steps in the Evercam journey. Product-market-fit is one thing, but you need your go-to-market. You've got to have a way of delivering it that is profitable. Identifying your customer profile, who you can efficiently sell to. There's all these pieces, and it changed over time. Ireland would be our home market, and now in Ireland we're very well established. It would be very unusual for there to be a large project that we're not involved with. UK similar, we're very dominant there. What took us global was the mission critical space, the data center, pharmaceutical, semiconductor, the big expensive projects that are very, very high value and where feedback is just extra important. If a problem is detected late, you're adding a lot of zeros onto the cost of that problem. Cameras are fantastic for spotting those issues early.

Roland Siebelink — Host (07:49)
So that's actually the monetization dimension of product-market-fit then, understanding how your product actually lets a customer save money.

Marco Herbst (08:03)
In our industry, construction tech people listening, they'll be very familiar with: are you selling to the contractor, are you selling to the owner, are you selling to subcontractors? And it really matters. Sometimes you end up having to sell to all of them, as we do. Construction projects have an end date, so as a business you want predictable revenue, we want long term commitments. We don't want all this churn at the end of every project. Getting that part of your model right is really significant. And construction is very regional. One of the big transformations that happened is — if we talk about early stage to scale up — was this idea around being regional. When we're selling into Australia, into Singapore, the Middle East, across Europe, Ireland, England, and America, each of those markets is very different and we're at a very different stage in each of those markets. I was finding myself in a board meeting trying to explain our numbers to my board of directors, but having massive complexities between what was happening in Australia, America, and the Irish-English market dynamics. To try and tell a coherent story was very hard. The big change we made to our business was dividing it up into business unit profit centers, so regional profit centers that report up a P&L.

Roland Siebelink — Host (09:57)
You hired a CCO in March. Was that your first CCO? Were you running founder-led sales all the way until then?

Marco Herbst (10:08)
Yeah, I've been very lucky in the Evercam story in that I've always had a business partner who has really looked after the commercial side of things. From the very beginning, Vinnie Quinn is my co-founder, also with whom we did jobs.ie together. Vinnie was always the customer facing end, which allowed me to talk to the programmers and the machines, which is my comfort zone. At a certain point, as we started to go into UK and Europe, I got to meet Randall O'Sullivan, who was just a fantastic — if Vinnie was the zero to one co-founder, Randall was the one to 10, 1 million to 10 million, which is pretty much exactly actually what Randall did do for us. And now as the company is just going beyond ten million, I got to meet Paul Devlin, who was really familiar with platforms, really familiar with that 10 to 100 period.

Roland Siebelink — Host (11:08)
What would you say is different in people's thinking when they're more a zero to one person versus a one to 10 person versus a 10 to 100 person?

Marco Herbst (11:24)
Zero to one, I think you're chasing a lot of things 'cause you're still looking for product-market-fit. Vinnie grew the business from that point of view. He discovered construction, discovered those customers. He had to be much more of a generalist. He had to understand more of the tech. He didn't want to get too constrained to any particular system. Randall really understood the sales floor. He wanted to take Vinnie's success and turn it into a system. He gave us the first structure on our sales team. The first playbooks, the first specializations of a BDR team versus the sales team versus customer success. If I think about what Paul is doing for us now, it's a whole level of maturity beyond that. It's not just revenue, but it's the quality of revenue. How predictable is that revenue? How far in the future is it? Really understanding the pipeline, what Paul calls revenue intelligence. Value conversations change. We're generally selling a little further up the org. We're talking more about trust and positioning and strategic alignment with our customers.

Roland Siebelink — Host (13:27)
Tell me a little bit more about how AI has been changing reality driven intelligence for you.

Marco Herbst (13:32)
It's unbelievable. We were always interested in what could be done with the power of the image. And computer vision has been certainly talked about for a very long time. In controlled environments, it has worked over the last few years. But construction is probably the least controlled environment you'll ever work in. In the last 12 months, what we're now getting from images and video with the VLMs is extraordinary. It would have started with relatively simple object detection. Our first product four years ago was counting the vehicles as they arrived on site. In hindsight, relatively straightforward. It happens to be very interesting for our customers because it's a strong indicator of is the project on track or not. But now we're much more interested in detecting hazardous situations, people working at height, people working with hot objects. These are much more subjective observations.

Roland Siebelink — Host (14:44)
For liability's sake or safety reasons?

Marco Herbst (14:44)
Yeah. The essence of construction safety is detecting hazardous activities and preparing for them. Most hazardous activities are pre-planned. There are supposed to be risk assessments done in advance. So if those things are happening on site and the risk assessments don't exist or they were unscheduled, that's a thing. And it's a thing that we want to report upstream, often to some of our software partners who are also being used by our customers. Companies like Navatech or HammerTech in the safety space.

Roland Siebelink — Host (15:15)
Can you tell us a little bit more about the transition from contractor sales to owner sales?

Marco Herbst (15:26)
As we got established and as it became better understood, the value of the project management type cameras that we sell, owners started to realize that this is data that they wanted control of and they wanted to run. Sometimes 'cause of disputes, but also sometimes just to have that independent verification of what was going on. And they are the ultimate beneficiary. When a project is delivered on time and on budget, the owner wins. It was a natural evolution. Now that we've been at this game for 15 years nearly, we have a massive database of stories from our customers, actual anecdotes. We've turned them into 100 of the most commonly quoted workflows that deliver value from reality driven intelligence. We literally have workflows.evercam.io, and it's a list of 100 use cases/workflows where reality capture is delivering monetizable value. I find out from you about your upcoming project. You're building a data center in Ohio, it's a $180 million project, it's going to take two and a half years — I've seen that before. We feed it into the system. I know what workflows are going to exist on your project because it's the same on the other projects. And I'm able to put money values and time values on it. I can tell you with a good level of confidence exactly how much you'll make from the project. That's new. That's really, really unheard of. And it moves us away from this insurance defensive based product to something where the customer can see this is a transaction that's gonna make me money.

Roland Siebelink — Host (19:30)
You're just putting 100 most common workflows openly on your website. Why give that away?

Marco Herbst (19:52)
One, because ultimately I think it's the right thing to do. But we are in our heart a company that believes in open. My life experience as a fan of technology and the Internet and what we're doing now is sharing information leads to good outcomes. Evercam at its very best, we're running the workflows, the agentic workflows that are feeding our customers' systems. The wider and broader this ecosystem is, the better. The faster it's adopted, the better. And I fundamentally believe if we are open and sharing what we call the Evercam 100, that is a tide that will lift all boats. Yes. Even if it didn't make business sense, I'd probably do it anyway. But we are of the belief that it does make business sense and we're of the belief that over the long arc, collaboration wins.

Roland Siebelink — Host (21:03)
Let's talk about your own evolution as a founder, Marco. You mentioned something in the prep call that you've gone through a quite tremendous evolution all the way from being the font police to management by joy. Can you describe your evolution?

Marco Herbst (21:32)
I think every founder hopes that at some point they might find themselves part of a global category leader opportunity. Having that size of opportunity in front of you sets a few rules — like the talent you need, the capital you need, the speed you have to execute at. If I go back in time 15 years ago, as we were still trying to find our way and not run out of money, it was a different mindset and it was much more controlling. Now, as we're all experiencing, we are seeing these incredible new technologies that are just changing how things are being done in ways that are very hard to predict. I describe it as the great acceleration because I don't think there's anything else that I can definitively say. Internally, in Evercam, what I do care about is that we're paying attention and we are all trying as best we can to engage with these new tools. I've actually found when people are coming to me with, shall I try this or shall I try this, my question back is which option would seem more fun? Which would give you more joy? It's seeming to work, funnily enough. Management by joy is the new management by constraint.

Roland Siebelink — Host (23:47)
Where did you grow up? What were you like as a kid, as a teenager?

Marco Herbst (24:10)
I'm basically here speaking to you from my home, my childhood home, here in beautiful County Wicklow. And I was always difficult and precocious and that maybe hasn't changed. The entrepreneurial thing — I'm surrounded here by a lot of my father's creations. He was a serial entrepreneur in the classic sense. He has an agricultural business here and a software technology company. My choices were either go work for him or chart my own path. And we even still today we would struggle to boil an egg together, so it had to be option B.

Roland Siebelink — Host (24:58)
Has he also been an inspiration for you as an entrepreneur?

Marco Herbst (24:58)
Very much. He's a great advocate for thinking differently and thinking critically. He's very, very slow to accept any sort of authority. And that also inspires one to question and want to get back down to fundamental principles. It's that systems thinking of wanting to piece together the story on your own terms that I think is really helpful.

Roland Siebelink — Host (25:33)
What would you tell founders is the most important thing you want to think differently about?

Marco Herbst (25:34)
Again and again and again, I just think people overestimate the downsides of things. For most of our lifetimes, we've lived in a world that collaboration and openness was the given direction. We're living in a slightly different time period at the moment where I think there's a lot more concern about the future. It's a lot less obvious that we're all going to get along. I think it requires more effort these days to re-rationalize the positives, the belief that actually, one plus one can be three, that by working together we will have better outcomes. I just believe that is scientifically true. It is worthy of reminding ourselves of that. I believe that's still to be true. It just requires a little bit more effort to look beyond the headlines.

Roland Siebelink — Host (27:21)
I love that. That's a perfect way of finishing the podcast. Marco Herbst, the CEO and founder of Evercam. Thank you so much for being on the show.

Marco Herbst (27:49)
Thank you, Roland. Much appreciated.

Roland Siebelink — Host (27:51)
And for the listeners, of course, we'll have another founder on the show, Breakthrough AI Operators, next week. Please do keep tuning in.