BBVA internationalizes its artificial intelligence factory to grow its AI product range

BBVA AI Factory researches the latest global trends in artificial intelligence and develops analytical components to enhance the bank's diverse technological solutions, such as the financial health functionalities of the BBVA app or the personalized recommendation of financial products. With the opening of two centers in Mexico and Türkiye – the regions with the largest volume of customers and business of the Group, together with Spain – BBVA internationalizes this factory to promote the creation of its products and services on a global scale and the transformation of its internal processes with AI.

To staff the new centers, BBVA hired 230 new professionals in advanced analytics and artificial intelligence, who now join the 200 at the Spanish headquarters. New hires include data scientists, machine learning engineers, developers, data architects and digital product experts. BBVA AI Factory thus doubles its talent specializing in data and artificial intelligence to more than 400 professionals.

The three centers are already driving the bank's various AI developments in a coordinated effort. Some projects are globally synchronized from the outset, while in other cases local development is undertaken in order to adapt a solution to the rest of BBVA's markets. The project design process is coordinated jointly, enabling BBVA to fully leverage the diversity of knowledge and talents across the Group. Some of the projects BBVA AI Factory is currently working on, hand in hand with BBVA's technology and business teams, involve the creation of a fully conversational virtual assistant that will simplify the user experience in digital channels, or the development of 'copilot' tools that will help managers to increase the efficiency of their work and improve the customer relationship model.

"We are fully aligned with the needs of the business areas. We develop our analytical components to address them, which allows us to transfer the maximum value from AI to customers," explains Ignacio Teulón, Global Head of BBVA AI Factory. "But, at the same time, our cross-cutting approach allows us to see how we can reuse these 'lego pieces' in other, entirely different business challenges. In this way, we optimize the work we do."

Ignacio Teulón points to two main lines of work: on the one hand, the transformation of the banking business: "AI gives us a deep understanding of our customers. With this insight, we can offer them authentically personalized financial health management solutions." Secondly, the AI unit engages with internal transformation, which involves equipping the bank's different units with tools that enable them to be more productive and generate new prototypes more quickly, such as generative AI and the alliance with OpenAI to introduce ChatGPT Enterprise in BBVA's internal processes. "Artificial intelligence gives us even more powerful means to continue placing the customer at the center of our strategy, but it also helps us approach problems from a new perspective," says Teulón, whose more than 20-year career at BBVA has helped drive its digital transformation, make the leap into mobile banking with its internationally recognized app and, now, reinforce the Group's AI strategy.

A strategy based on reusing code

Code reuse is one of the pillars on which BBVA's artificial intelligence developments are based, as it shortens product creation and implementation times in all markets.

BBVA AI Factory developers centralize their components in the Mercury library, which currently has more than 500 pieces of code. On average, around 100 BBVA data scientists use it every month. The resources available in Mercury help create many of the data-driven functionalities that the bank makes available to customers. Among them, the bank transaction categorizer, expense and income forecasting or the detection of anomalous situations in the account. This code is also used internally to evaluate the performance of data models, such as a component that has been used to bring transparency and explainability to the 'machine learning' models with which the bank detects and performs early debt management, which the publication Global Finance has awarded with one of its The Innovators 2024 awards.

In 2023, the factory shared part of this library with the open source community through the collaborative platform Github. The goal is to contribute to the innovation of the fintech ecosystem and evolve Mercury with external contributions. Every month, an average of more than 900 downloads of this open library are produced on Github, which other developers can use in their own projects.

As future challenges, Ignacio Teulón highlights the need to continue creating analytical components that respond to specific business challenges while maintaining reusability potential. In addition, global best practices in advanced analytics that can be transferred to local businesses should be shared. In addition, Teulón considers it crucial not to lose the balance between applied development and trend research. The bank must seize the new opportunity presented by generative AI to continue to offer added value to customers and optimize internal processes.