As part of the Crédit Agricole Generative Artificial Intelligence (AI)* managed adoption strategy, the Group DataLab (ITD) has developed CA Generative Search. The aim of this application is to offer users a new natural language research experience on a wide range of documents specific to the Crédit Agricole businesses, supplying well-drafted, relevant responses based on easily verifiable sources.
Rigorously assessed on a corpus of public documents, its results can be further improved by supplying Generative AI with less complex and better structured documents. This solution has been applied with CACIB on its “Augmented Sales” project, based on its market research and internal servers. Numerous other businesses hope to benefit from it soon.
The Group R&D DataLab is also working on internally controlled LLM** to improve response accuracy and reduce roll-out and usage costs (invoices based on text length including question, context and answer), as well as the consumption of resources inherent to commercial Cloud services.
This programme is part of the Crédit Agricole Group’s Societal Project which strives to act in favour of the climate and achieve carbon neutrality by 2050, while strengthening its position as a stakeholder in the societal and environmental transition.
Indeed, in addition to the necessary securitisation of sensitive data when used on the Cloud, the environmental impact of Generative AI, which is highly energy intensive, must be considered in light of the Group's commitments to sustainable IT services. |