" class="no-js "lang="en-US"> EXCLUSIVE: "The AI Factory" - Jesús Renero, BBVA in 'The Fintech Magazine'
Sunday, April 14, 2024

EXCLUSIVE: “The AI Factory” – Jesús Renero, BBVA in ‘The Fintech Magazine’

Spanish bank BBVA has just set up its second technology hothouse as part of an ambitious tech strategy that will also see it hire thousands more IT professionals in 2023. Head of Advanced Analytics Jesús Renero talked us through it

As far back as 2017, BBVA, Spain’s second-largest bank, stared long and hard into a crystal ball and saw its future as a data-driven technology company.

Investing heavily to produce in-house IT capabilities and artificial intelligence-led solutions, BBVA had, within two years, opened a development hub it calls the AI Factory near its central headquarters in Madrid. There has been no slowdown in investment since. A second AI Factory has recently been established in Mexico; the bank is keen to see if it can reuse analytical models to accelerate the roll-out of AI-based products across multiple countries at the same time.

And the bank is aiming to hire more than 2,600 extra technology professionals this year. By the end of June, it had already recruited 1,200 of them, adding to the 22,000 STEM professionals it employs globally, which represent about 19 per cent of the bank’s workforce. More evidence of BBVA’s determined technology drive came last year when it formed a global software unit, bringing together more than 16,000 developers, to accelerate the creation of digital solutions for the bank and their scalability among countries where it operates.

“One of AI Factory’s core values is to embrace open innovation,“ says Jesús Renero, head of advanced analytics at the hub. “We have created a cross group inside the Factory, which is coordinating every experimentation that we need to do; to A/B test whatever idea we have that will face the final user, the customer

“In that sense, innovation has to move together with experimentation and we think it needs to be bottom-up. We don’t have a dedicated team for innovation. Instead, we identify from different programmes something that is valuable, and worth the time experimenting with. Then we propose that idea, and, if agreed, we start experimenting and innovating with that. That is the way we understand innovation in this area.”

The way the Factory goes about its work is governed by four initiatives that foster exchange of knowledge and ideas. Every fortnight it holds The Discussion Club, a one-hour forum that is open to all, where technical challenges are shared, feedback sought, and discussions held about different points of view when approaching a project.

“Innovation has to move together with experimentation and we think it needs to be bottom-up. We don’t have a dedicated team for innovation “

Then there is what it calls a Quick Study, which involves a small, theoretical investigation into a new technique that could be applied to solve a problem or
challenge and can be shared among teams. Anyone interested in the issue can participate, whether it directly affects the team or just to learn more about the topic.

Thirdly, its X Program aims to develop functional prototypes to provide solutions to shared challenges, which will be used by different teams. The goal is to create a new workspace to experiment, implement and test the initial phases of a novel product idea, based on the data. Fourthly, time is set aside to experiment with state-of-the-art technology in Innovation Sprints. They allow work on ideas that could be further developed and included in BBVA’s AI-based solutions.

These serve as the seed of X Program. But BBVA is also unafraid to look outside its own organisation to help it provide the best for its customers, Renero stresses.

“We try to cooperate as much as possible with the universities that are around us. We exchange knowledge very often, and run small projects with them. These do not involve complicated agreements in terms of data, because we don’t believe that exchanging data between organisations is the way to go,” he says. “We also try to keep an eye on open innovation, where the open innovation organisation within the bank – called BBVA Spark – helps us to identify potential startups that could help us with specific problems. We cooperate with open innovation organisations in some other universities, too.”

The net result of all this work is that, by the end of 2022, BBVA had increased its strategic data projects seven-fold. It aims to manage more than 500 initiatives by the end of 2023.

For example, the bank now has 12 financial health app functionalities that were developed in the AI Factory, including services for automatic classification of expenses into different categories and the visualisation of digital subscriptions, be they streaming services, or utility supplies.

These financial health services generate up to 40 personalised alerts for customers including, for example, warning of a higher-than-usual bill, accounts going into overdraft or offering suggestions for saving and creating a financial cushion. BBVA’s financial health tools are already available to customers in Spain and they are being rolled out in Mexico, Turkey, Peru, Argentina and Colombia.

Proving their business worth, they are currently consulted by 18.5 million unique customers (July 2023), 38 per cent more than in the same month in 2022, and registered almost 93 million interactions, 93 per cent more than in July the previous year.

A REASONED APPROACH

Another area of focus is sustainability. The team has developed several tools that show the energy consumed, the costs associated with the mobility of customers, and the carbon footprint created by individuals and companies. A sustainable car calculator allows customers to see the costs associated with an electric car, versus one powered by fossil fuels. And it’s now using geopositioning technology to work out how much money individual customers could save by installing solar panels on their homes. And then there is risk management.

“Instead of detecting fraudulent events, we work on what is the normal behaviour and try to identify those things that we are not able to predict “

“We are exploring ideas that are probably common ground for most people, but in the banking industry, and under the regulatory issues and frameworks, we think they are valuable and worth the effort,” Renero says. “One of them which allows us to come up with the best probable solution to any risk management problem proposed by the bank.

“The other big area for risk management is causality. Causality in risk management allows us to understand the reason behind the behaviour of different users. This is pretty new, the technology is still not mature, but we think that there are many opportunities in that area, as well.”

Fighting fraud is one of the huge challenges facing all areas of financial services. Continuing its spirit of cooperation, BBVA has recently joined forces with Banco Santander and CaixaBank to create FrauDfense, a project that will enable the exchange of relevant information and data in order to prevent financial crime. The alliance will first develop a secure tool to share information on fraudulent practices and effective response measures. Other banks and companies from other industries interested in exchanging information on fraud to protect customers, entities and broader society will also be able to join.

BBVA is also experimenting with other, innovative AI models to detect and prevent fraud, as Renero explains: “We explore deep learning models to treat the landscape of movements and transactions of a specific customer as if they were images – things that the neural network is able to consider, or to model more effectively. Customer behaviour is roughly predictable. So we consider every transaction movement and event like a time series – we can say there is a given likelihood that something will happen in a given period of time.

“So, instead of detecting fraudulent events, we work on what is the normal behaviour and try to identify those things that we are not able to predict. Whatever is outside of that expected behaviour is what we call relevant fact. It’s relevant because we couldn’t predict it. That is something that we flag to the customer, or communicate to the bank to take a look.

“That is one area where I think that we are experimenting successfully on how to improve the current rates of detection.”

BBVA prides itself on following a mantra of ‘smart and responsible use of data’. Underscoring that ethos, Renero says that all the AI models it develops go through a peer review process to assess the impact on people’s lives and are subjected to explainability and interpretability studies.

It will be interesting to see how that develops in relation to generative AI. Early this year, the bank struck a deal with Amazon Web Services to use the global platform’s Cloud-based machine learning and analytics at the very heart of its data strategy.

Renero says there is ‘tremendous potential’ for generative AI, but also cautions that the technology is not yet fully mature.

“We believe that the technology is useful, and we need to explore that – we are working on understanding its full potential. Ultimately, we believe that, in three years’ time, a significant number of models in our portfolio will be based on large language models.


 

This article was published in The Fintech Magazine Issue 29, Page 21-22

People In This Post

Companies In This Post

  1. BNY Mellon Collaborates with Accenture to Drive Accelerated Growth for Clients Read more
  2. XS.com Awarded the “Best Multi-Asset Broker – LATAM” at UF Awards in Mexico Read more
  3. Insurtech bolt Appoints CFO, Matt More, SaaS Technology Veteran Read more
  4. Navy Federal Credit Union Announces 7-Year Strategic Partnership With Backbase Read more
  5. aelf Leads the Fusion of AI and Blockchain to Shape the Future of Technology Read more
Global-blockchain-show