BNP Paribas has integrated generative AI to significantly reduce the time staff spend on preparing ESG (Environmental, Social, and Governance) assessment reports.
The ESG Assessment serves as a vital tool for BNP’s relationship managers, enabling them to evaluate clients’ ESG performance and associated risks. This task, previously taking several hours, involves answering up to 80 questions that address client risks related to governance, business ethics, climate impact, biodiversity, and human rights.
To enhance efficiency, the bank developed an internal model utilizing ChatGPT and Microsoft Azure. This model scans and indexes documents provided by clients, subsequently suggesting answers for the questionnaire based on the uploaded materials. Users can easily access relevant sections of these documents through linked comments.
Enam Ehe, head of ESG data and systems, stated, “Since the launch of the AI Assistant, it has been utilized by 600 users, facilitating nearly 1,600 ESG Assessments. Currently, the AI Assistant contributes to approximately 35% of all ESG Assessments, a figure that continues to grow. We assess the AI’s effectiveness through both quantitative and qualitative monitoring, in addition to tracking the operational costs of the LLM model in Azure, which remain minimal in this context. Overall, user adoption has been strong, and feedback has been predominantly positive.”