On June 5th, more than 200 participants gathered at the headquarters of BGL BNP Paribas for the third edition of the Luxembourg Analytics Summit, organized by SAS, in collaboration with the Luxembourgish bank. During an inspiring session of talks, local and international experts took the stage to discuss the latest trends and use cases in the field of AI.

Marc Aguilar (Chief Data Officer at BGL BNP Paribas) officially opened the summit: "Today, it is possible to develop an algorithm in 8 weeks, but infrastructure and data sourcing strategies are needed. Also, quality data is required to work on developing an Artificial Intelligence". According to the CDO, data is the new asset which must be managed through the use of data quality management tools. It requires governance and a sourcing strategy: where does the data come from? Who do we buy it from? BGL BNP Paribas has opened a lab center that is currently working, with a startup, on an AI-powered solution aiming at detecting fraud. The bank is also focusing its efforts on the use of Machine Learning when it comes to authorizing payments. For instance, if a customer has multiple bank accounts, but one of his payments cannot be processed because a couple of euros are missing on one of these accounts, a tool gathering all his data might help to evaluate to risk and use the money on another account. "At BGL BNP Paribas, we decided to work on applying AI to improve operations. Implementing new technologies takes time, and, according to us, it is easier to measure the ROI on operational efficiencies," Marc Aguilar added.

Master of Ceremony Rein Bryssinck (Sales Director at SAS Luxembourg) then took the stage and highlighted the fact that every single industry, work function and aspect of our personal lives will be affected by Artificial Intelligence. "It will notably replace from 20% to 40% of today's jobs, while creating new ones," he added.

Mr. Bryssinck welcomed his colleague Michel Philippens, Head of Customer Solutions at SAS, who delivered an inspiring speech entitled "Artificial Intelligence: Separating the hype from reality". He started: "My first encounter with AI was the TV show called Knight Rider, which was created more than 40 years ago: a self-driving, intelligent car. And nowadays, most of its features are being produced".  He then shared a definition of AI and explained that it can now be applied to a wide and diverse set of human tasks, from blocking fraud, processing payments, help with loan applications, etc. "But the most powerful tool, behind AI, is analytics," added Mr. Philippens, who continued: "AI can make automated decisions or support complex human decisions". There are several examples of companies that work with SAS to leverage AI and data to create new services. SciSports, for example, is a company which gathers data on soccer players to assess their performance. Then, the experts resell these insights to professional clubs. Michel Philippens also mentioned the fight against colon cancer, as "AI has the power to change healthcare". SAS is currently working with the Amsterdam Cancer Research Center to analyze the appropriate time when treatments should be proposed. On the marketing side, AI will act as a relief for customer service and aims at predicting the intent of the customer. "One of the best promises of AI is also that it will facilitate the green lane: it will make sure safe customers can pass through a green lane, making the service easier and faster," Michel explained. As a conclusion, he said that AI offers a myriad of possibilities but that professionals should try to find a platform to cover all technology capabilities.

SAS and BGL BNP Paribas then welcomed David Hagen (Head of IT supervision & Supervision of Support PFS at CSSF) who focused mainly on the use of AI in the financial sectors and the regulatory challenges it implies. "When it comes to AI, sometimes, the human is in the loop, sometimes he is not. Therefore, finding a unique definition is complicated as the scope is very large," started Mr. Hagen. Nowadays, many solutions such as RPA, chatbots, KYC intelligent agents, text recognition, credit scoring, etc, are powered by AI. The CSSF expert also highlighted the difference between ethics and governance: "ethics is about what is acceptable to do with AI (the goal), while governance is about how AI should be used (the means)".  According to Mr. Hagen and the CSSF, outsourcing is one of the most important points as it must be mastered by regulators and financial institutions. He then listed some of the challenges brought through the use of AI: project management and costs, change management, etc. He concluded: "AI is disruptive for financial institutions and understanding the main issues of AI is key for the board of directors. Moreover, AI should not lead to a loss of control over the business. Concerning regulators, they should also benefit from AI in the months to come. It is a true game-changer in terms of governance, risks and control."

Professor Benny Mantin, University of Luxembourg, then shared his knowledge on AI applied in the field of logistics. "At the Luxembourg Centre for Logistics and Supply Chain Management, we have one moto: innovate, impact and inspire," said Prof. Mantin, who also insisted on the quality of data: "Basic AI seeks to refine an answer based on data provided, if it fails rules need to be redefined. Machine Learning imitates the human brain, while Deep Learning uses multilayered neural networks to facilitate learning." In the field of logistics, Benny Mantin defined three big categories and departments that can benefit from AI: management, marketing and operations. More precisely, it will have a huge impact on supply chain, namely network planning, inventory management, warehouse/distribution operations, etc. "Route optimization, thanks to the rise of electric and automated vehicles, will also be improved," added Prof. Mantin. All things considered, there is a possibility to reduce logistics costs by at least 5%. "AI brings massive change but also many gains. It requires companies to rethink their business processes and models. Routine jobs will diminish in the years to come, while new ones will emerge. Also, regulation is needed," he concluded.

After the break, the organizers welcomed Jos Polfliet (Head of Machine Learning at Faktion) who shared many AI use cases with the audience. He started with three pieces of advice when thinking of applying AI to your business: assess what is going on in your company (talk to each other, define the issues you face, etc), build a strategy telling you where you want to go (find a good use case, from call center automation, employee churn analytics and cybercrime detection to cost analytics, lead scoring and customer segmentation), and finally execute to go from A to Z. "Yet, there is only one use case: solving problems," highlighted Mr. Polfliet, underlining the fact that using AI is only relevant if used to solve a business problem. And actually, there are problems and inefficiencies to be solved in every single department and industry." Also, AI can help saving costs and selling/producing more". To resume, two major strategies for finding use cases are important: the first is looking for problems and trying to map them to mathematical statements that can then be translated to Artificial Intelligence. The second strategy is taking use cases that are known to produce ROI and seeing if they apply to your company.

"Will your next CMO be an algorithm?" was the name of the presentation given by Albert Derasse, Managing Director & Partner at Inbox. In a context of digitalization and with the advent of AI, the role of marketers is changing rapidly: they must clearly review their missions and upgrade their skills. "The introduction of new technologies, multichannel development and more demanding customers makes it difficult for traditional marketers to thrive. But now, they can add intelligence to their data aiming at the hyper personalization of customer relationship. Algorithms can help," he explained, before adding: "Marketers need to accept the fact that they cannot understand everything and that they have to rely on technology. It is a big revolution for them". According to him, tools are embedded in the future success of the marketing strategy and therefore the entire company. "Marketing won't be driven by ego anymore, but by acknowledging that experts also need technology, while still using their own curiosity and creativity," concluded Albert Derasse.

Finally, Denis Batalov (Principal Solutions Architect at Amazon Web Services) took the stage for the last presentation of the day and he talked about the history and democratization of Artificial Intelligence technologies. "Amazon.com has been using a recommendation system for more than two decades, and the company introduced an app called Amazon Remembers as early as in 2008. Users could send a picture and we would find the product and send a link to buy it," he stated. He then shared several examples of AI-powered operations, such as robots helping with the logistics, and the famous Amazon Echo and Alexa, but also the work of the e-commerce giant on autonomous cars. Denis Batalov ended his presentation by entertaining the crowd using Amazon's face recognition systems. He concluded: "If there is one message to remember, it is the following: using AI, you should expect inaccuracies and errors. Be prepared to handle them".


Download the e-book “Making Sense of AI” to learn about the boundaries of AI, as well as the many ways that modern AI applications can improve our understanding of the world and enable us to make better, faster decisions. 


Alexandre Keilmann

Photos: Dominique Gaul

Publié le 22 juin 2018