“The Impact of Artificial Intelligence Is Also Societal”
Sylvain Duranton,Global Leader for BCG GAMMA interviewed by Florent Vairet (Les Echos Executive)
[Florent Vairet]How has a Frenchman ended up at the head of BCG Gamma?
[Sylvain Duranton]I joined BCG twenty-five years ago, for my first full-time job. After a few years in BCG offices abroad, I returned to Paris, where I created the company’s pricing activity, and I joined the Paris office’s Executive Committee. On the strength of my long history at the company, I was asked to create BCG Gamma, the global entity dedicated to data science. So, yes, it’s true, when I was appointed, many people told me: “Only at BCG can a Frenchman be made boss of an entity like that.”
But I should also say, without false modesty, that it was quite natural. When it comes to technological subjects, the French do have a good press. France rode the first waves of artificial intelligence in the 1970s, and the country’s mathematics education remains highly regarded. The French have always been well represented at BCG, and the Paris office has a special place in the organization: it acts as a kind of laboratory in the network.
Artificial intelligence is a very technical subject. How do you demystify it?
When I graduated from the Ecole des Mines and Ecole Polytechnique at the end of the 1980s, data science was not really a subject that they taught. We manipulated data, but it was more a question of applied mathematics.
Nevertheless, having this background and, especially, learning computer coding were crucial to enabling me to demystify algorithms. I have been working in these fields ever since, and I now devote 100% of my professional energy to them. I have read a lot and taken many training courses. However, my aim is not to be a practitioner of AI. My job is above all to put AI at the service of companies.
Have any people in AI made a particular mark on you?
Raymond Kurzweil could be one. He compared the functioning of the human brain to that of algorithms, and he showed that there could not be any meaningful direct comparison. Machines do not “understand” anything, in the human sense of understanding. A machine is very fast, whereas a human brain is very slow.
However, our organized and conceptual way of thinking is extremely effective. It is important to reassure people and to make them understand that we are not trying to replicate the human brain. In France, the member of Parliament and Fields medalist Cédric Villani, who was our scientific adviser before his election, is promoting a calmer view of AI that is going in the right direction.
Major foreign investments have been announced in France. Is it a problem that we have no national leader in AI?
In this domain, we have to think on a European scale. Everyone wants to have a great champion in their own country, but the first priority is to make sure that French companies take AI on board to become more competitive. If a European or French solution emerges, that’s great, but let’s first concentrate our efforts on rethinking companies’ organization and operating methods, and make sure that we attract top talent.
At BCG Gamma, the number of employee has increased from 25 to 400 in three years. What exactly can you promise to entice the best?
We offer them ambitious projects. We try to recruit the most respected data scientists in their domain, with, at the very least, a master’s degree and often a PhD in data science. In this very fierce battle for leading talent, we are competing with very attractive startups and with the GAFA companies, which provide the chance to work on enormous systems.
At BCG, it is not the salaries we can offer that really make the difference. We promise our prospective new talent that although they will be working on a small scale, the projects envisaged are incredible. For example, anticipating the side effects of treatment for multiple sclerosis, constructing a unique optical signature for diamonds, or creating a decision assistance system in real time for the control rooms of major airports. Our engineers work in small teams on 6- to 12-month projects.
This year, we will be organizing 2000 recruitment interviews, and we plan to hire 200 people to sustain the growth of this activity, which more than doubled between 2016 and 2017.
What type of manager are you?
I attach immense importance to helping employees grow. The more they can do, the happier they are. You have to build an ecosystem with your employees, where they take on more responsibilities than they could ever have imagined.
Do you have a management model?
No, not really. I mistrust great charismatic leaders. I often tell myself that at the foot of a big tree nothing much grows. I attempt to cultivate the increasing skill levels of my teams, and I hope that when I leave my job at BCG Gamma, people hardly notice, because the succession will be fully prepared to take over.
Doesn’t the workload scare people off?
I don’t think so. Everyone we have recruited has a passion for data science. Often, they even code for their own pleasure on weekends on Kaggle, a web platform that organizes competitions in data science.
So, for you, what comes next?
At the moment, my aim is to continue to nurture the growth of BCG Gamma. At the same time, I am active in civil society. I am the “Straight Ally for LGBT Network” at BCG, a position that I use to combat homophobia, which still causes a lot of suffering in western societies.
I am also treasurer of the Aurore association, which has 1,500 employees and is committed to working alongside people in situations of economic instability and social exclusion to increase their autonomy, via housing, health care, and occupational integration. Finally, I am also treasurer of Lycée Saint-Louis-de-Gonzague, where my children go to school.
Why take on so many commitments?
BCG has helped me grow, and I have acquired skills that I am now utilizing on the outside, by helping NGOs and associations manage their funds and define their strategy. I donate skills. In return, these voluntary activities nourish my professional life. The impact of artificial intelligence is not just economic. It is also societal. You cannot act solely as a technician. You have to engage with society in order to understand the impact of the transformations that we are undertaking. ■