How Capgemini is working on AI adoption in banks, financial institutions

The financial system continues to spearhead the adoption of productivity technologies, such as AI and generative AI (genAI), including in Latin America.

However, regulatory and compliance issues, along with a lack of clarity and definition of scope, still permeate many of these implementations.

That is the opinion of two top Latin America analysts at French IT consultancy Capgemini who have worked with various financial institutions in the region to build business models and implement AI tools.

“AI at more traditional banks won't be something immediately at scale and focused on customer processes. There's still a lot of 'tall grass' to cut through in their back-office,” Daniela Dutra, head of solutions for payments and banking at Capgemini, tells BNamericas.

Dutra refers to the areas of management, internal processes and systems. The executive sees two fronts emerging, where the risk involved is lower and the ease of AI implementation is greater.

They are smart document management, or the use of AI to scan and interpret reports and logs, and the use of AI tools in a co-pilot model, as assistants for agents in the back-office. “We've seen many cases in this latter field, banks have been asking us for this a lot,” she said.

Next comes what the executive calls intelligent customer contact, moving from the back-office a little closer to the front-end, with genAI powering chatbots and other channels in the relationship with the consumer.

Processes related to money laundering, investment advisory and identity proofing/customer profiling are other areas in which financial institutions are beginning to apply more advanced and automated AI tools, she said.

Wealth management, in particular, is a promising area for the use of such technology, according to Dutra. 

In a recent report, Capgemini said that "by integrating behavioral finance with artificial intelligence, global wealth management firms can better recognize and address HNWI [high-net worth individuals] client needs; and genAI can aid hyper-personalization of relationship manager/client experiences and communications."

INTEGRATION

Whether in the deep back-office or in areas with greater customer contact, there is still an important challenge of integrating the new platforms with the management of processes, people and legacy systems.

“Integration is the main gap we're seeing. This is what will make AI scale up. If it's being used in isolation, it's not even a differentiator, it's simply another tool, a prototype,” she added.

Daniel Berman, director of business consulting at Capgemini Brasil, says that it is necessary to distinguish "ready-made" innovation from "complementary" innovation. 

A ready-made innovation, he says, would be a car, which has a single function, a specific defined purpose. In contrast, electricity would be a complementary innovation, a means for various other functions and processes. 

“In this sense, AI is a lot like electricity. What's more, AI is still very young. Electricity became ubiquitous and scaled with uses developed with it over time, from public lighting to connecting homes, which also reduced its cost. The same will tend to happen with AI,” Berman told BNamericas.

Berman sees it as imperative for institutions to clearly define use cases where AI can be helpful and make a practical difference when applied. For now, he sees genAI as good for replacing recurring and repetitive work of a lower complexity.

RISKS

Capgemini, like other consultancies, works with clients to structure projects and define suppliers.

In the case of banks, before embarking on genAI adoption projects, all Capgemini clients created internal governance areas to establish guardrails and guidelines in the use of technologies, say the executives.

“There are suppliers and suppliers. We have massive, open AI systems without much security, and we have companies and suppliers that are already starting to utilize these solutions for regulated, specific, personalized use, and following specific guidelines,” said Dutra.

In addition to governance, the Capgemini executives highlight the importance of automated observability tools. In other words, tools that orchestrate and oversee the functioning of AI.

“If you need a human to observe the AI – not to make use of it – you’ve gone two steps back,” Dutra said.

LATIN AMERICA

In Latin America, Capgemini sees Brazil as most advanced in the adoption of genAI and other technological tools in the financial system due to its advanced bancarization and the process of financial digitization, which is more widespread than in other markets in the region.

Because of its proximity and close links with the US market, Mexico is next most advanced in the region. In both cases, relative economic stability is also a big factor, they say.

“Mexico, in this field, is the Brazil of five or six years ago,” said Berman, referring to formal inclusion of the public in banking and the state of digitization.

According to Dutra, 15% of Mexico's population is now banked, which is reflected in opportunities for the sector.

The recent launch of Brazilian digital banks and fintechs in Mexico, such as Nubank, Banco Inter, and C6, are examples of this development, she said.

Finally, Colombia is a market where Capgemini has worked a lot with clients in the financial sector and which has become more open to new regulatory arrangements and new payment models, they underscored.

“All in all, Latin America is fertile ground for these technologies. Until then, there is still a lot of experimentation and testing and a lot of things to happen," said Dutra.