Making a difference in people’s lives and empowering them to make the best decision regarding their finances is the number one drive for a transformation in financial services – or at least it should inspire companies that intend to generate a more significant impact. Years ago, the first spark of this revolution was ignited by smartphones, which triggered the first wave of Fintechs, putting a computer in each person’s pocket.
But what will come next? As many of you may imagine, open banking in Brazil unlocked a new chapter of this revolution and has the responsible use of data as a critical component.
The role of Artificial Intelligence at Nubank
However, along with a myriad of fresh possibilities, open banking also presents new challenges. Having access, according to regulations, to data from different financial institutions is useless if you cannot fully comprehend them, extracting the signals from the noise.
That’s where artificial intelligence and machine learning come in. They support innovative companies to clean, standardize, and classify all transaction history to later generate insights about customer financial patterns and journeys for their benefit.
Almost nine years ago, Nubank was the pioneer in building superior digital products and experiences to fight complexity through transparency, and ultimately redefined what it means to be a bank customer. Now, we see nearly limitless potential for generating value in this new paradigm also based on transactional data intelligence.
Customer onboarding, credit decision, and fraud prevention are near the top potential experiences to be re-designed, which also embrace creating a self-driving customer experience and improving customer service, operations, security, cross-sell, up-sell, and beyond. All that while staying true to our commitment to data protection and striving for the best AI ethics practices.
And why are these new services so relevant in our current landscape?
It is a fact that most Brazilians struggle with their finances. Sixty million of them, representing 40% of the adult population, have bad credit (nome sujo no Serasa, as we say in Portuguese). The scenario worsened during the Covid-19 pandemic, with 52% saying they were stressed about their expenses and financial commitments.
Only 21% say that they always or often save money, while 44% save rarely or never do it, according to the Serasa: Mapa da Inadimplência e Renegociação de Dívidas no Brasil research.
That said, machine learning is an invaluable tool for scaling credit products responsibly while we take the time to understand customer pain points and aspirations, educate them about how to address those, and craft financial solutions for them.
Artificial intelligence can help us achieve the same customer experience at scale via customization, product recommendations, and tools for positive habit formation and financial coaching.
Finally, we aspire to have an AI-driven ecosystem for financial wellness that is as well-developed and helpful as the consumer health and wellness ecosystem is today – taking into account the highest standards of privacy and data protection, and keeping in mind the aspirations and needs of those who benefited from our services.
Data revolution can foster customers’ autonomy
In simple words, imagine a bank able to, among other things, predict liquidity, illiquidity, and spending events to give personalized offers, advice, and experiences. Imagine an app that, before your payday, offers you to invest a percentage of it based on your potential savings over this period, or before you overdraft, your checking account due to a cash flow gap, offer a short-term personal line.
That is just a small sample of how the data revolution can foster customers’ autonomy and guide them through their own information to lead them to mindful decisions.
At Nu, we are beginning a journey toward extracting meaning from banking transactions by applying distinct AI and ML techniques. By using supervised and unsupervised machine learning models to classify these transactions, we aim to provide relevant insights so that machines and customers can understand and act on the information.
An example: Daniel shopped for groceries
Let’s use an example of a typical credit card transaction: the customer Daniel shopped for groceries. Our models first separate the transaction description into substrings that represent information such as the payment processor, merchant name, store location, and the number of installments.
Then they clean, categorize and enrich these small pieces of information by searching in our knowledge graph for the words and concepts associated. And based on that, machines predict future transactions and make personalized recommendations so that Daniel can spend better and save more on his future grocery purchases. As a result, we leverage the power of high-quality data to support customers to enhance their daily financial decision-making.
At the end of the day, customers must always be at the center of our business decisions, no matter the technological revolution we are passing. We now have the opportunity to create a self-driving mobile experience tailored to each customer’s needs in real-time. It means responding to customer events as they occur and enabling developers to build quickly, create, train, and deploy machine-learning through lighter and faster infrastructure.
The outcomes of this perfect match are just around the corner. Are you ready for this next chapter?