Transcript of the interview:
Antoine: Today, as part of our Innovation Leaders podcast, we’re very fortunate to welcome Nicolas Tomatis, Director of BlueBotics. Nicolas, we’ve known each other for… I did the math: 29 years!
Nicolas: Wow, we’re getting old!
Antoine: It’s a real pleasure to have you here today. I’ve always followed your professional journey with great admiration. To start, could you tell us a little about your career since your studies at EPFL?
Nicolas: My studies at EPFL date back to 1998. I then completed a PhD in robotics, which ended with a great experience we shared together: developing that famous robot for Expo.02.
That project eventually led to the creation of a company, which officially took it over and brought it to the exhibition. After the six months of Expo.02, we found ourselves in a tiny office with eleven or twelve of those mobile interactive robots that we had developed with EPFL and our teams.
There were basically just the robots and four employees wondering: “What are we going to do when we grow up?”
From there, we analysed the market and realised that the navigation technology we had developed could be relevant to the industrial market, particularly for the automation of industrial vehicles.
That’s how the idea emerged: to bring disruptive technological innovation to this market. Then, of course, we still had to find a way to bring it to market.
Antoine: So when BlueBotics was launched, you were finishing your PhD. Did you immediately take over as Managing Director? It was a very small team back then, but it has grown considerably since. How many people are you today?
Nicolas: Today, we’re around sixty people. We’re part of a group and we also have offices in the United States and China.
But I actually got into all of this somewhat by chance. I wasn’t the director at the beginning. I was one of the people developing the technology and, in reality, nobody was really in charge.
We thought democracy was a good idea! Then we had some coaches who explained to us that democracy wasn’t necessarily the best way to run a company and that we needed to define a leader.
So we had to choose one. There were four of us, and I don’t know why they pointed at me, but that’s how I ended up taking on the role in 2003.
Antoine: It’s a great story. Congratulations on everything you’ve achieved. Before we go into more detail about other projects, I’d like to hear your view on innovation. What does innovation mean to you today?
Nicolas: The term “innovation” is used everywhere and sometimes for just about anything. The basic definition is something new — a positive novelty.
But I prefer a definition that isn’t mine. It comes from Scott Berkun, who wrote The Myths of Innovation. He essentially says that you can call something an innovation once there is a customer willing to pay for it.
I think that’s quite an interesting definition.
Antoine: Yes, I like that one! Now, looking at your journey, you’ve experienced innovation at a university, then in a start-up with just a few engineers, and today in a company of around sixty people.
How do you maintain that ability to innovate as the company grows, the number of customers increases and the challenges become more complex? I often talk about keeping a certain “childlike spirit”. How do you see it?
Nicolas: That’s one question containing quite a few others!
I like this idea of a “childlike spirit”. What immediately comes to mind is the quote often attributed to Mark Twain: “They did it because nobody told them it was impossible.”
At our company, when we face a challenge, we first try to generate our own ideas without immediately looking at what others have already done.
I think that’s an interesting approach because as soon as you start by looking at existing solutions, you risk being channelled into what has already been done.
Once we’ve generated our own ideas, of course we do our research, look at scientific publications and review the state of the art. But keeping that childlike ability to come up with new ideas is fundamental if you want to innovate.
Then, regarding our evolution, you have to understand that there are two types of start-ups. There are those that identify a problem and use technology — innovative or not — to solve it in a new way. And then there are those that have an amazing technology… that doesn’t actually serve any purpose yet, so they have to find an application for it.
Unfortunately — or fortunately — we belonged to the second category.
We started with a technology that was genuinely disruptive compared with what existed on the market at the time. Back then, industrial vehicles were guided using cables embedded in the floor, magnetic strips or, for the most innovative systems, large reflective targets detected by lasers.
We came along and used the laser sensor that was already there for safety — the one that detects the surroundings to prevent the vehicle from hitting someone. We used that data to create a map of the environment.
This already existed in research, but not really in industry. So we industrialised the technology to bring much greater flexibility to automated vehicles and make their installation significantly easier.
We started with a disruptive innovation for which we then had to find an application. After that, things change: you have to convince customers and open doors in an extremely conservative market.
We even had to hide some of our innovations because they were simply too much for certain customers! Obstacle avoidance, for example — the ability of a machine to dynamically react to its surroundings — was something we had to hide for seven years because it was too innovative for customers. We simply weren’t credible.
Antoine: Incredible.
Nicolas: And then things change again. You enter the market, start gaining customers and build credibility. Those customers have needs that you have to address.
Innovation then becomes more incremental — step by step — in order to keep satisfying your customers.
And that brings us to what Clayton Christensen calls the Innovator’s Dilemma. On the one hand, you have to continue with incremental innovation to meet the needs of your existing customers. On the other hand, you have to make sure that someone else doesn’t arrive with a disruptive innovation and take your place.
So you need to find the right balance between the two. At BlueBotics, we try to invest energy in both directions.
Antoine: I really like that point about the adoption of disruptive technology. In the specific case of BlueBotics, can you quantify how long it took for the market to adopt your technology?
Nicolas: Yes, unfortunately, I can!
We founded the company in 2001. It took us around two years to find our direction and identify the market we wanted to target — industrial logistics with automated vehicles.
Then it took us roughly three years to find a customer crazy enough to trust us. That customer was highly innovative himself, very much one of those early adopters who love trying new technologies.
We started working with him in 2006, and the first installation took place in 2009.
From 2009 to 2012, we were finding somewhere between zero and one new customer per year and installing fewer than ten systems annually.
Then, around 2013–2014, volumes started increasing, particularly because some customers were selling more automated vehicles equipped with our system.
In 2014, two or three important references gave us much greater visibility. From there, things really accelerated, and we started gaining six, ten, fifteen, twenty and eventually thirty new customers per year.
But it remains an extremely slow market. From the day we sign with a customer to the point where they bring their own product to market, it takes an average of around a year and a half. And that’s just for their first vehicle.
For us to really start making money, they then have to sell their products and purchase more of our systems. So the whole process takes years.
Antoine: Today, we almost have to talk about artificial intelligence because it’s everywhere. Do you use AI in the algorithms behind your autonomous navigation systems? Is it something that could make things easier for you in the future? More generally, what’s your view on AI today at BlueBotics?
Nicolas: That’s another complex question!
Until three or four years ago, I had actually banned the word “intelligent” from our company because we wanted to stay away from all the hype surrounding artificial intelligence.
You have to remember that for the past fifty years, roughly every decade, there has been a huge wave of excitement around artificial intelligence that arrives… and then fades away.
Today, it’s quite interesting to see that almost every robotics start-up includes “artificial intelligence” in its positioning in one way or another. It’s no longer simply robotics — it’s Physical AI. It’s wonderful! That’s what you have to do today if you want to attract investors.
In our specific case, the technologies we’ve been using for the past twenty years could already, in some ways, be considered artificial intelligence technologies.
But the current wave is mainly linked to the new language models, which can genuinely bring value.
We’re also using and testing these tools. For example, we use them to improve the speed or quality of the texts we produce, and we’re testing systems that can automatically propose unit tests when creating code.
There are many tools out there that we’re currently evaluating. We’ve already adopted some of them because they allow us to work more efficiently.
However, I still believe that when you’re working at a very high technological level, today’s tools have limitations, particularly in terms of the complexity they can actually handle.
It’s also important to remember that current artificial intelligence systems don’t truly understand what they’re giving us. It’s impressive because they can produce extremely convincing sentences, but the system itself doesn’t actually know what it’s telling us.
So we have to be careful about how we use these technologies. I think we still overestimate what we can expect from them today.
Antoine: I share that view. There are obviously incredible time savings. At the office, we use AI for research, for example, or when we encounter fairly complex bugs, particularly in low-level programming. After spending a few hours investigating something, AI can sometimes help us find a solution — provided we ask it the right questions.
Nicolas: We’ll have to exchange ideas about these tools because I’m curious to hear about your experience. That’s precisely where it becomes interesting — being able to share different perspectives on how these tools can be used, particularly for expert-to-expert programming.
Antoine: Throughout your professional career, whether during your PhD or at BlueBotics, you must have faced a real technological challenge at some point. Is there one you could tell us about and explain how you overcame it?
Nicolas: Well, because I was a bad engineer, they quickly got me out of engineering and made me a manager!
More seriously, I’ll come back to what we were discussing earlier. I’ve tried to establish a culture within the company where we start from a naïve perspective and only afterwards look at what exists in the literature and in research.
We first develop a naïve concept. Then we evolve it by looking at existing references and the state of the art on the market. Finally, whenever possible, we work directly with experts.
We present our idea to them — under confidentiality if we’re considering patenting it, for example — so that we can confront our concept with their expertise.
We really try to go through these three stages before committing to a development project that could require a significant investment.
But ultimately, the biggest challenges I’ve faced within the company have been less technological and more commercial.
The real challenge was finding the match between this amazing technology we had developed — which initially had no real purpose — and a market.
We had to find the right business model, successfully bring the technology to market, convince customers and generate growth. That’s where a huge amount of work was required.
Antoine: I still think that having been an engineer and having taken part in the development was a real strength throughout your career. Even though you haven’t been directly involved in development for a long time, it must be easier to present a product to a customer when you understand the technological challenges behind it.
Nicolas: I think it’s extremely useful, provided you also know how to step out of the engineer’s mindset. Because if you stay entirely within it, it doesn’t work.
But it’s true that I’ve been lucky enough to do more or less every job within the company.
If I remember correctly, until around 2007, I was still vacuuming the office before customers arrived!
Then one day, when I changed computers, I asked my colleague Rémy to install the development tools for me. He looked at me and said: “No, I’m not doing that. You shouldn’t be developing anymore. You need to do your job.”
And actually, he was absolutely right.
Antoine: Good old Rémy!
There’s one question I definitely have to ask you in this podcast. We’ve worked together, you knew FiveCo before it even became FiveCo, you knew the five founders, and we’ve also worked on several projects for BlueBotics.
I’d like to hear your outside perspective: how do you see FiveCo? How are my colleagues and I perceived?
Nicolas: For me and for my colleagues who know you, it’s quite simple.
Whenever you want to work with a supplier, you always ask yourself several questions: Are they reliable? Will we be able to work well together? Will communication be easy?
The advantage we have with you is that we know you and we’ve already completed several projects together.
We have absolutely no doubts when it comes to quality, organisation or meeting deadlines. We don’t even question it.
That’s also why you probably work more with the people I refer to you than with us directly! Because FiveCo is genuinely a reference for us.
If someone tells me, “We need an engineering company,” I tell them: “Go to them.”
Well, that guy can be a bit stubborn… but apart from that, everything works very well!
Antoine: Thank you very much, Nicolas. To close this interview, we ask all our guests the same question. And I’m going to ask you to stick to the rule: define innovation in just one word.
Nicolas: I’ll give you a short phrase instead, if you allow me: value creation.
And if you really want just one word: value. But that might be a little less clear.
Antoine: I really like that. We’ve heard a lot about risk recently, but I think “value creation” is very accurate. In a way, it also implies risk.
Nicolas: Yes, it goes hand in hand. But if I’ve done everything I’ve done — even back when I was in research — it was because I was looking for that value creation.
That’s also one of the reasons I left research. I realised that research was only the first step. I wanted to have customers who were genuinely happy because we were delivering something useful to them, and because they were willing to pay us for something they believed had value.
So, for me, value creation really is the definition of innovation.
Antoine: A huge thank you for joining us today.
Nicolas: Thank you.
Antoine: We’ll include all the details and further information on the BlueBotics website, which is very comprehensive and provides a great overview of the technology. Thank you very much.
Nicolas: Thank you.