The banking jobs : CTO - Artificial Intelligence Lab BNP Paribas CIB
Ludan Stoecklé became interested in Artificial Intelligence (AI) while a student at INSA Lyon. He worked 12 years in AI focused start-ups, and then joined the Data Science & Artificial Intelligence Lab of BNP Paribas CIB. As CTO, he oversees the design and development of AI software for the bank's internal teams. In this interview, he shares more about his business, the Group's strategy, and the new challenges facing banks in terms of AI.
Tell us about your role in BNP Paribas CIB's Data Science & Artificial Intelligence Lab.
To give you a bit of context, the Lab is a team that develops AI projects before industrializing them: an internal software publisher. These AI solutions are then available for our internal clients, either within CIB or the wider BNP Paribas Group. As CTO, I have a cross-functional role across all AI topics, including: Translation, Speech-To-Text, Image Content Analysis, and many more. My personal expertise is in Natural Language Generation (NLG) and Chatbots, and I support Lab teams at every stage of project development, from design to production. We also contribute to defining and implementing the Group's strategy for AI, anticipating the company's technological needs, and identifying opportunities to innovate.
You have been working at BNP Paribas for a year, now. What is your career path? Why BNP Paribas?
I studied Computer Science at INSA Lyon, where I started to get interested in AI. A topic that, at the time, was much less popular than today! I then worked for more than 12 years in AI startups. I was one of the founders and technical director of Yseop, a vendor of Natural Language Generation software, before joining Addventa, an AI services company that operates primarily in banking and finance. I chose to join BNP Paribas a little more than a year ago, because it is the only European company with a strategy of building in-house AI solutions and that has secured important resources to implement it. I was also seduced by the scientific and technical level of the Lab teams.
Artificial intelligence is revolutionizing the financial sector. What are the new issues for banks?
There are several considerations. On the one hand, we are facing the emergence of new competitors, including GAFA, who are gradually entering the banking market and positioning themselves within traditional value chain activities, such as online payments - with services such as Google Pay and Apple Pay. On the other hand, the competition of online banks is particularly strong, since they massively use AI to improve the user experience. Therefore, we have to be just as agile and innovative as they are to stay up to date.
The last issue is the automation of all repetitive and basic tasks not requiring human intervention. Specifically, the idea is to allow employees to refocus their time on what they can do best. At BNP Paribas, for example, AI frees up time for our operators in the back, middle, or front office so they can be fully available for their exchanges with clients, performing complex analyses and making decisions. This pushes us to ask: “How can AI improve on what humans do best?" or "What organizational changes are needed to accommodate this new form of human-AI collaboration?"
the competition of online banks is strong, since they massively use AI to improve the user experience. we have to be just as agile and innovative as they are to stay up to date.
What is BNP Paribas' strategy for artificial intelligence? What is happening with other banks?
BNP Paribas wants to preserve its independence and competitiveness over the long term. To do this, the bank has favored a "build" approach: this means that we develop our own Artificial Intelligence bricks, train them on our internal data, and host them on our own internal cloud. The idea is to take full advantage of AI, like all other banking players, while maintaining a high level of control over the data and the decisions taken by the AI software.
The idea is to take full advantage of AI, like all other banking players, while maintaining a high level of control over the data and the decisions taken by the AI software.
Our belief is that we cannot trust a black box developed by an Internet giant, and hosted on a foreign cloud, to make decisions for us. However, many other banks have a different strategy, called "buy": they rely on external AI specialists and vendors. We are one of the only banks in the world that has made the choice to create our own internal AI software publisher. Though this strategy has only been in place for a few years, it has already proven successful allowing us to develop at the same speed and with the same quality standards as external vendors, often at lower costs, while acquiring and developing AI skills internally.
Specifically, the idea is to allow employees to refocus their time on what they can do best.
Could you give us examples of AI projects that you and your teams are currently working on?
A project that we are very proud of is our translation software we call "Translate": similar to Google Translate, but entirely developed and hosted internally, and much better on text using financial vocabulary. We are currently adding a feature to translate an entire document (PDF, Word) in one single step, and in just a few seconds. Our current users are eagerly waiting for this feature! Another example is our search engine, "Search": it is currently in production for CIB and we are working to deploy it at the Group level. Similar to Google, it indexes all sources of information that employees have access to, and allows them to instantly search for internal information. Finally, we are developing AI solutions that perform document analysis. The "Document Intelligence" project aims to extract structured information from our and our clients’ documents to ease the analytical tasks of employees.
What are you the most proud of since you arrived at BNP Paribas?
Recently, we set up a platform that allows the creation of chatbots in an industrial way, quickly and at a low cost, for any team wishing to automate its response process to frequently asked questions received from other internal collaborators. A few weeks ago, we promoted this solution on our intranet, and more than 50 teams around the world contacted us to express their interest in this service! Most are small entities that receive recurring internal requests, but have neither the time nor the computing resources to develop sophisticated automatic response tools. We help them create tailor-made chatbots that meet their specific needs. I am very proud to be able to help them concretely, efficiently, and quickly.
Last word: what is your next big AI challenge?
I have one particular AI ambition in mind: contributing to existing open-source components and publishing new ones developed by BNP Paribas. Today, we freely use open-source components published by the community to develop our solutions. In return, we want to put some of our components open source too. In addition to helping the community, this will be a testimony of BNP Paribas’ leadership in AI. I hope we can achieve this in the coming months.
How would you explain your job to children ?
I would say that I am like a teacher. My role is to teach computers to perform operations currently reserved for human beings. I teach them, for example, to recognize an image, read and understand a text, or answer a question. In fact, the Lab is a kind of elementary school (or Pre-K!).
What is your morning routine when you arrive at the office?
First, I choose a desk with the best view of the outside, because it is very important for me to see a piece of nature when I work. Then I take a coffee and once awake, I catch-up with my colleagues.
Sandwich or canteen?
I love going to the canteen, the value is just extraordinary! If you have the opportunity, I recommend giving it a shot…
Personal office or open-space?
It depends. I use my personal desktop when I work from home. However, when I am in the office, I use the open-space and take advantage of the freedom offered by the flex office to choose a desk "with view", either of the sheep in the meadow or the ducks in the canal.
Crédit photo : header ©araraadt // ©metamorworks // ©kokotewan