Artificial Intelligence And Fraud Detection: which opportunities?
Thierry Derungs, our Chief Digital Officer, was invited to speak at Vivatech about Artificial Intelligence and fraud detection. Here’s his take on this hot topic.
AI is in fact nothing new. Companies have been trying for a long time to improve their efficiency, speed and accuracy through automation. The first foundations of deep learning even date from 1965! What has drastically changed is our affordable computing power and the accessible volume of data, which are the two fundamentals of AI.
The 2 simple letters A and I hide a completely new world, full of opportunities. Every day, brilliant minds share and/or patent new models, approaches, languages… while also making them easier and more accessible.
I am fascinated by this fast-evolving world, which embed not only AI but also other new technologies such as edge computing and blockchain. Nevertheless, although I believe that digital has never been so human-related, I often wonder about our capacity to change in order to embrace these new opportunities. For instance, if today’s User Experience on tech devices is very fluid, people often forget that behind this sensation of ‘easiness’ lies a great deal of complexity. So, without expecting everyone to develop a high tech profile, it’s necessary to open up our employees’ understanding of these new technologies, for it to be less subjective.
Listening to many consultancy companies or next-door CEOs, you often hear two statements:
Just put some AI box and the magic will take place. Easy and effortless!
If you do not have AI, your business will be blown away. Don’t you want to survive?
So… I see many enterprises, not only banks, jumping onto this topic with ‘fight-for-survival’ expectations while dedicating too limited resources (the ‘plug & go’ illusion).
My presentation was to share our experience at BNP Paribas Wealth Management, as indeed there are many opportunities with AI. However, they request a well-balanced step-by-step approach, thus allowing a good learning curve that goes fast-enough.
Security and fraud detection have always been a non-negotiable must-have at BNP Paribas. We have a strong responsibility towards our clients and everyone in the company, from our IT staff to our top management, is highly committed to preserving the security of our clients’ assets. Through our startup ecosystem, we saw the promising capacity of a behavioural approach in detecting fraud. Not only when a fraud actually occurs, but also when it’s emerging at its earliest stage.
By combining behavioural analytics together with machine learning , we can build new individual detection patterns while having the capacity to adapt individually. Let me give you an example: although you and I might be alike, we remain different in many ways. So, if we were to be both included into one same pattern, the trick is nonetheless to notice (and take action) when our habits differ and change. As such, we keep learning on an individual basis, thus strengthening our ability in terms of speed and accuracy to detect and prevent fraud.
Working with the startup NetGuardians has allowed us to focus on our own risk expertise while making the most out of AI’s capabilities. With NetGuardians, we were able to achieve a very rich and clear case management allowing our investigators to further analyse the suspicious cases.
Change management, in terms of soft skills, must never be neglected nor underestimated. Especially when the pace of new technologies goes much faster than our usual human adaptation capacity.
I truly love these two quotes from Einstein: “It’s harder to crack prejudice than an atom.” And
“God gave me the stubbornness of a mule and a fairly keen scent”.
Awareness, change management and clear explanations based on business use cases are the best way to move from a “No” to a “Mmmh” then a “Why not” and finally the expected “Let’s try”. If the business use case has been carefully selected, the collaborative approach to deliver a Minimum Viable Product will provide the final incentive to test such an approach.
Any five-year prediction is very risky considering the fast-evolving pace of AI and new technologies. Of course, I could vacuously state that AI will be key. We all know that.
What is trickier to predict is for what (and how far) AI will be used to support our clients’ promises and fulfil their expectations.
What is certain is that we have a business vision. We constantly challenge ourselves and, for each new initiative that we launch, we involve our clients from day 1. From the conception of the Minimum Viable Product to its use in production, we collect their feedback. Based on that, we adapt. AI, together with other new technologies, will surely be a key asset to support our client engagement.
Thierry Derungs is driven by passion: his family, work equity - white, black, male, female, young, old… "who cares when the job must be done, leadership, international collaboration and, of course, all things digital." He believes in human beings and their incredible power when they work together.
After more than 15 years at ING, Thierry Derungs joined BNP Paribas in 2009. Having successfully led a wide range of digital projects and programs for the past 20 years has given him a strong foundation in international management, strategic business planning and program management.
As Chief Digital Officer, he develops and oversees the Digital Strategy for Wealth Management. Whether implementing new web and mobile solutions for our clients, or making fundamental improvements to our employees’ digital work environment, he firmly believes that each and every initiative is an equally important step along the path to bring you our value.