Brest, 26 September 2024
Crédit Mutuel Arkéa is currently testing a chatbot based on generative artificial intelligence with 500 members and customers. The aim is to capitalise on innovation to boost operational efficiency, with high-quality customer support available 24/7, while freeing up time for bank advisers so that they can strengthen their role in advising their customers.

Drawing on its expertise in data, artificial intelligence and generative AI, Crédit Mutuel Arkéa wanted to offer an initial panel of retail customers a chatbot using a solution based on generative AI. This initial test will enable customers to interact, from their personal space on the website, with a conversational chatbot, which will instantly answer their questions rather than redirecting them to the FAQs already available on the bank's transactional website. ‘Generative artificial intelligence is now widely available, both in the workplace and in our daily lives. It is an undeniable source of opportunities. Before offering our customers solutions based on generative AI, we felt it was essential to gain a deeper understanding of how it works in order to control the risks involved,’ explains Elisabeth Quellec, Head of Crédit Mutuel Arkéa's Retail Customer Business Unit.

Crédit Mutuel Arkéa's chatbot covers 600 questions

To carry out this test, Crédit Mutuel Arkéa is using the technology offered by Microsoft Azure, based on the GPT4-o model developed by Open AI. A multidisciplinary team, made up of staff from the Marketing and Customer Experience Department, the Customer Operations and Advisers Department and the Group's Innovation and Operations Division, worked on this project to apply the generative AI solution to a corpus of 600 frequently asked questions. This analysis has enabled us to define a set of topics on which the chatbot equipped with the generative AI solution will be able to provide appropriate answers in a conversational mode.

Enhancing the value of the banking advice offered by the local mutuals

The aim of the approach is to improve the reachability of customer support and reduce the number of unqualified calls. Crédit Mutuel Arkéa is also responding to the changing expectations of its customers, , who want more autonomy in terms of navigation, with the possibility of finding immediate and precise answers to their questions themselves. When requests require more specific expertise or advice, Crédit Mutuel Arkéa's chatbot will encourage customers to contact a bank advisor to arrange a meeting. The aim of this approach is to make the most of new technological solutions while enhancing the banking advice offered within the local mutual network.

Innovating in the responsible use of personal data

The testing of this chatbot with generative AI therefore serves Crédit Mutuel Arkéa's long-term ambition to innovate in the use of the data at its disposal to better understand its customers and enhance its value proposition. Elisabeth Quellec: ‘This innovative solution is in line with our commitment to the responsible use of data. We are convinced that this chatbot equipped with generative AI will meet the new needs of our customers, particularly in terms of contactability and autonomy, while strengthening the role of our banking advisers within our network, thus combining proximity and innovation to serve our territories."

Further information

Crédit Mutuel Arkéa, which via the Data Office of its Innovation & Operations Division has obtained the ‘Trusted AI Methodology’ certification issued by the Laboratoire National de Métrologie et d'Essais (LNE), is also a member of the ArGiMi consortium. The consortium won the call for projects entitled ‘Digital Commons for Generative Artificial Intelligence’, launched by Bpifrance as part of the France 2030 initiative.

In October 2023, Crédit Mutuel Arkéa also implemented a generative AI model designed by its in-house teams in open access for developers on the Hugging Face platform. This initial work has enabled the teams to gain operational expertise on the subject of generative AI, particularly in relation to the issues of sovereignty and energy sobriety posed by this type of technology.