AI in banking: a revolution in the making

Banking sector transformation: accelerated by challenges, leveraged by technology

On March 19, 2024, Club Les Echos Débats “Tech & Innovation” in partnership with Wavestone, Celonis, and Rennes School of Business welcomed Sophie Heller, COO of BNP Paribas Commercial, Personal Banking & Services Group, Mariam Rassai, Chief Digital & Data Officer of BNP Paribas Wealth Management Group and Catherine Mathon, CIO of BNP Paribas Leasing Solutions Group, to discuss “AI in banking: a Revolution in the Making”, moderated by Charlie Perreau, Head of the Startup section at Les Echos, and introduced by Johann Chazelle, Associate Partner at Wavestone.

Banking sector transforms around 4 main AI challenges:

1. New customer lifestyles
2. Performance targets for retail activities in a context of rapidly rising interest rates
3. New “Beyond Banking” service models for mobility and housing

4. Regulatory and CSR requirements

AI is key to meeting these challenges, and banking players have been exploring this technology for almost 10 years. Today, nearly 90% of large companies consider data and AI to be a priority for their development.

Increasing investment has highlighted limitations in access to data, technology industrialization and the need for scaling up specific business lines.

Generative AI is game-changing, with unprecedented adoption and potential across all banking activities, particularly customer relations, IT or operations.

5 AI challenges for banking to overcome:

1. Clear expectations from business lines
2. Pursued IS transformation
3. Readiness for Generative AI
4. Risks and regulations
5. Employees onboarding

Choice of AI solutions (in-house vs. market, including startups)

BNP Paribas pursues a hybrid strategy, with:

• Market solutions: AI services are sometimes directly integrated to the software (as in the case of inbound call management).

• Tailor-made solutions: For customer data processing or to adapt models to enterprise data.

Note: the BNP Paribas Group’s strategy is focused on the independence and sovereignty of its infrastructures, which is reflected in the solutions selected.

BNP Paribas’ AI organization

A Group Data Office defines the strategic guidelines and provides regulatory guidance. In addition, each business unit and country has its own organization, sharing information globally about:

• Ideas (use cases…)
• Methods (scoring…)
• Solutions (document management…)
• Code

An inventory of all the use cases, called the “AI Watch Tower”, has been deployed. It embeds a description of each use case deployed and a common methodology for measuring performance and value creation (A/B testing).

Beyond technology, how human can it be?

All three speakers unanimously agreed that, for BNP Paribas, people are at the heart of customer relations, and AI will remain at the service of people.

Certain tasks are bound to disappear, and new skills to be developed (around prompting, for example). Meanwhile, AI will take the load off some activities, so that advisors can focus on their core business: supporting customers.

The BNP Paribas Group invests heavily in the cultural transformation and onboarding of its employees, through:

• Leaders and teams training: 2-day Summer School for 2,000 employees, Master Class for Comex leaders.
• Idea and solution sharing: Brainstorms with diverse participants, communities capitalizing on experience.
• Agility: Enabling business and IT teams to work together.

AI compliance and safety

Artificial Intelligence is designed and implemented in line with BNP Paribas’ values and code of conduct, avoiding any form of bias. This begins with raising awareness of data and its uses.

Within a strict regulatory framework, confidentiality and security issues are essential, as is energy sobriety. In this respect, the use of public generative AI for professional purposes has been banned, and experiments are only carried out in secure environments, in compliance with the Group’s ethical charter.

The Group is moving forward in a reasoned and iterative way, to master compliance and safety issues “by design”, particularly in a changing regulatory context (such as the recent AI Act).

Generative AI: first use cases

BNP Paribas’ first Generative AI use cases revolve around content generation and document search/analysis/synthesis, for customers and employees.

• Leasing: CSR reports in accordance with circular economy; carbon footprint; etc.
• Retail: video generation for simple, customized communication about an individual’s savings or the portfolio’s performance; report generation or e-mail search via personal assistants for employees; etc.
• Wealth Management: search / navigation in documentary corpus and internal procedures.

Traditional AI use cases will benefit from Generative AI. For example, compliance checks carried out during phone order-taking conversations.

Women in Tech

BNP Paribas invests in mentoring from secondary school onwards, encouraging young women to join these challenging professions, which offer attractive jobs and the potential for successful careers.

One third of BNP Paribas IT employees are women. The “Women & Girls in Tech” meetings provide a platform for discussing their paths.