After years in Sweden, France and the Silicon Valley, Nicole now spends her time between Zurich and Berlin. As the CEO of Merantix Labs, she is also part of the Merantix management board where she actively shapes our ecosystem and advises entrepreneurs. Not least of all, Nicole is a renowned public expert on the topics of digitalization and artificial intelligence - affiliated with the Aspen Institute, the World Economic Forum and the University of St. Gallen.
Speaking to me in the midst of the Corona pandemic, we talked via Zoom - with many laughs and surprisingly few connectivity issues.
Finn: Nicole - you are the CEO of Merantix Labs. But before we talk about your company and your role, do you want to introduce yourself? What did you do before joining Merantix?
Nicole: I am an economist and econometrician by training. I basically studied old-school data, so to speak [laughter]. After working in finance for a little while, I segwayed into tech on a very specific game and auction theoretic ticket. My former boss [Paul Milgrom] actually just won a Nobel Prize in Economics. Don’t be afraid of niches, I think is the lesson here.
Finn: And then you founded your first company?
Nicole: Yes, in the data science machine learning space - even though I did not come from a computer science background. But the field is very dynamic and it is actually very important to also have people in the space from different fields - applications of the technology are spread just as widely as you could imagine, so it is really not a uni-dimensional space.
Finn: And how did you learn about Merantix?
Nicole: I have known Adrian for almost 20 years, I believe - since my first year in college. In fact, there are some really embarrassing photos of us: we all think we never aged, unfortunately though we do. And since then we kept having conversations about machine learning and what was going on in that space. To a degree, we actually built our companies in parallel. My own company in Switzerland did pretty much the same thing that Merantix Labs does today: machine learning projects with industry partners. But it was bootstrapped, no funding and it lacked the wider Merantix context.
Finn: And when you joined, it was clear that you would lead Merantix Labs? Did you join without a second thought?
Nicole: At the time I joined, there really was not yet this clear vision for what Merantix Labs could be. So, we worked on that in concert and I was able to bring in a lot of the experiences that I had made with my own company. It was not easy for me to leave my own company, not least of all since I am now sharing my life between two cities. But Merantix had a wider mission of building an ecosystem for AI and growing into a central AI hub within Europe. That’s what really made the difference for me.
Merantix Labs: Working on the future with industry partners
Finn: Let’s dive into Merantix Labs: what do you guys do?
Nicole: Most of my brothers and sisters at Merantix have product companies where they have a relatively clear path towards a specific product for a specific industry vertical. At Merantix Labs, we build bespoke AI solutions for companies. We take this deep knowledge and expertise across a wide field of domains in creating value from data and understanding data-driven business models - and bring it to already existing companies to build tailored machine learning products for them. In big law firms, e-commerce, warehouse management, publishers, chemicals - there really is a wide interest. We help to scale these applications and to truly harness the value of machine learning. A lot of organizations are struggling with this on a variety of levels: lack of talent, lack of experience, data and infrastructure issues.
Finn: How big is your team now?
Nicole: We are now close to 20 people. All are situated in Berlin of which two-thirds are engineers, and one-third are business team members. We work in interdisciplinary teams to help translate business problems into machine learning problems.
Finn: How much of your work is educational? Clearly, your team needs to understand what the company in question is doing - and how. And at the same time, it must be part of your job to explain the technology and what it could do.
Nicole: Yes, that is the constant stretch we are doing, no doubt. It is very challenging for our engineers to design and implement a solution that creates value and integrates well into the customers process. So, we have a lot to learn from industry and we can bring a lot of expertise to the company as well. It’s a constant translation process.
Finn: One of the first things I did when starting at Merantix one and a half years ago was to build a fund deck for our fundraising. I remember writing “consulting” when I did the Merantix Labs slides - and we had a longer discussion about this internally and subsequently stopped using “consulting”, but started to use somewhat artificial terms like “industry vehicle”. What is wrong with consulting?
Nicole: [laughter] Nothing is wrong with consulting! But I think by now people have pretty preconceived ideas about what consulting is - and usually the deliverable is a slide deck, whereas our deliverable is code. We ideate and really implement a solution customers can use.
Managing diverse teams
Finn: I asked Adrian what question he also wanted to ask you. He said: your team is majority female. How do you foster male achievement in your team? [laughter]
Nicole: I have thought about creating an ombudsman. [laughter]
Finn: But there is a serious point here: you seem to seek diversity actively and I am really curious to hear what is the main driver. Do you think diversity is good in itself? Does it have returns to business?
Nicole: Whether diversity is good in itself - that is a really interesting philosophical debate. But, yes - studies definitely show that diversity is a driver of excellence. And I would agree: it does drive more robust and better decision-making. But I’d also say it is important how you define diversity. For me, it is really not what you see in a picture but something that is reflected in one’s skill set, upbringing, social and economic background, mindset.
Finn: Agreed. But why does not everyone pursue diversity?
Nicole: Diversity has become this term that everyone uses and, in a sense, also romanticizes. While I agree it’s a goal to be strived for, we also have to be honest. Even though diverse teams have proven to outperform, they also require more skilled and empathetic management. Why do people like to manage their buddies from their golf club? Because they understand exactly how they work. In diverse teams everyone speaks a different language, literally and figuratively. So, it takes more effort and generally a more aware approach. But I think it is very much worth the effort!
Finn: So what is the secret to managing diverse teams?
Nicole: The first step in my view is accepting that people have different ways of working, tasks they draw energy from, motivation and also sensitivities. That’s why as a team we have a lot of sessions getting to know and working on understanding each other better. And then I really try to facilitate a culture of open and fairly explicit communication: you are reacting differently than I had anticipated - explain to me your point of view! Just a couple of weeks back, a team member held an internal session on The Culture Map where she introduced the book and we self-evaluated where we all see ourselves on these very crucial cultural dimensions.
But I also think it is important that you understand that you will never get it fully right, as everybody has biases and shortcomings - it’s human. You need to try very hard to alleviate these biases from an organizational design perspective and then be upfront about what you are trying to achieve: communicate openly and be straightforward when something went wrong. The good thing in leadership is that making the effort already gets you half-way there: people pick up on it and then you already have a culture that allows for growth - personally and as a team more generally.
Inclusive AI and the society
Finn: I am currently involved in a project that tries to enable a dialogue between AI industry experts and representatives of the civil society. The broader topic of the dialogue is Inclusive AI. Could you explain your understanding of inclusive AI and weigh in: should we think more about inclusivity? And only with regards to the design process or also about systems in production?
Nicole: We have a clear commitment at Merantix Labs to make AI inclusive. Thinking about it, there are definitely different layers. Who is actually creating the models and algorithms is already a super small percentage of the population. There, the question is how aware are they of the goals of the wider society. That certainly affects which use cases we are focusing on and where most innovation happens.
Secondly, there is a layer of who is empowered by the technology. As AI makers, we have to ensure that AI technologies benefit and empower as many people as possible.
Another topic is wealth creation. With the technology itself come properties that we tend not to think about a lot in the industry. If there is a tendency towards resource concentration in the hands of a few, then we need to think about a broader participation in the gains of automation. We are creating all these efficiencies and improving performances - but how do we want to split all this in society? So maybe I define inclusivity as trying to prevent the technology from being divisive…
Finn: The last point is really interesting. My personal intuition is definitely that the actual dividing line lies between those that know the technology, its potentials and dangers, and those that don't. And that is a problem. From a business perspective it definitely affects the willingness-to-adopt, but in the medium-run it might also cause political divisions and backlashes. I wonder sometimes how good of a job we have done in educating wider society about AI.
Nicole: A very poor job! The way technology more generally has been portrayed in pop culture and the media is 1) dystopian as 99 out of 100 films on AI end with the AI destroying manhood, 2) framed from a very passive perspective: it sounds like the technology is just there. But in reality it’s a software - someone is programming it and 3) creating expectations it can partly not live up to. I mean the autonomous car is still not really here even though it has been announced throughout the last 20 years. And in some ways, the pace is what really counts when you are dealing with enormous changes for society. We still have some way to go before we will see general AI taking over large parts of complex tasks.
Spending time and thought on public affairs
Finn: Last question: you are a very prominent figure in the European AI industry. Even as someone who has a fairly infrequent social media habit, I get the impression that you give a lot of talks, participate in a lot of panel discussions. Surely that eats away a lot of your time and mental capacity. What motivates you to work on public affairs?
Nicole: Since the beginning of technology the question has been whether it will empower the individual or be used to enslave people. I am a tech optimist and genuinely think AI can solve a lot of challenges we face - not only for individual businesses, also for wider society. And I want us to be proactive agents of our own faiths, so to speak, in Germany and Europe. And for that I am convinced we need to talk about positive examples, making things concrete, answering questions people have. It’s certainly a lot about taking responsibility as a tech-maker and then also facilitating adoption and some change, culturally and in the regulatory framework. And that is a great part of Merantix’s mission, too.
Finn: But surely it helps Merantix Labs to sell products and projects, too, no?
Nicole: No doubt about it. But that is also a part of this wider mission: creating business value and reaching people and sectors that are not yet immersed in the tech world. Take SMEs: they are the backbone of our economy. And that is not an abstract thing, right. It’s our jobs, the functionality of our welfare system, the stability of our society. At the same time, SMEs definitely have some serious structural disadvantages: siloed data, poor data quality, lack of resources. They carried us through the last industrial revolution and walking this journey now with them is incredibly fascinating and genuinely helpful. And for that you need to be visible.
Finn: That was very interesting! Thank you so much for joining me, Nicole. And: best of luck with Merantix Labs!