What is ESG Assessment? Who does it mainly impact and what role does it play in our ESG journey?
The ESG Assessment is the tool for Relationship Managers (RMs) to assess the ESG performance and risks of our clients. The ESG Assessment is implemented in the impACT platform, (the Group’s internal platform dedicated to ESG data).
ESG Assessment aims at incorporating the ESG parameters into the risk assessment of a client through a sectorial questionnaire and a review of their ESG controversies, both completed by the RMs (complemented by CSR team’s opinion in some cases) and it is fully embedded in the credit risk process. So far, the ESG Assessment has been deployed on Strategic clients (Segment 1), Large corporates (Segment 2), relevant Commercial clients (MidCaps & SME’s, Segments 3 and 4), on FIs (Banks, insurers, asset managers and sovereigns) and is progressively deployed in the Commercial Real Estate, ERI Project Finance and Aviation.
The ESG Assessment questionnaire allows capturing the salient risks of the sector analysed and enhancing the dialog with clients on their ESG Strategy, risks and opportunities.
As the regulation and ESG data risk management become more stringent, having a robust set up enabling customers’ ESG Assessment is key, while remaining competitive to serve our Business in an agile mode to reduce the time to market and offering stable and secure platform.
The ESG questionnaire is one of the main pillars of the ESG assessment process aiming at identifying the ESG profile of the customers by collecting substantial data in various ESG fields, such as governance and business ethics, climate pollution and biodiversity, and human rights. It has been deployed in 2021 for LOD1 (Relationship managers, CSR teams) and LOD2 (RISK, SCO, SBO) populations to identify the ESG risk parameters, inserted in a synthesis sheet used during the credit committee for credit granting decisions.
How does AI improve the accuracy and efficiency of the ESG Assessment?
As of today, it takes hours for a relationship manager to fill in the ESG Assessment, which can contain up to 80 questions. Using AI aims at helping reduce questionnaire filling time and gain operational efficiency. In fact, the main goal of implementation of Generative AI is to assist Relationship Managers in completing the ESG Assessment questionnaire faster and more efficiently, by providing suggestions answers and justifications based on quick comprehensive analysis of the documents disclosed by the clients and uploaded by the Relationship Manager. It also allows reducing data quality issues in the ESG Assessment completion in ImpACT by identifying and suggesting the right information to answer the question. Besides, AI makes the job simpler and quicker for the Relationship Managers, who remain responsible for the inputs provided.
"Besides, AI makes the job simpler and quicker for the Relationship Managers, who remain responsible for the inputs provided."
What data is precisely analysed by the AI?
All you must do is first upload the relevant ESG/CSR documents, disclosed by the clients, publicly or shared with the Relationship Managers during client meetings. Once the documents are uploaded, you can now launch the AI analysis. What does it mean exactly? Documents are being scanned and indexed by the AI, and the AI displays suggested answers for the questionnaire from the documents uploaded, and comments referencing the relevant sections of the document, that you can consult by clicking in the related links. AI is prompted and delivers as many answer suggestions as possible.
All the suggestions aim at helping the Relationship Managers in completing the ESG Assessment, especially to guide and ease the analysis of the client’s ESG strategy, to help identifying client’s main ESG risks and impacts and easily display the main sources of information and their location.
How and with whom was the tool developed?
Project was initially launched by CIB Sustainability Office and managed in collaboration with CPBS Company Engagement and many stakeholders from the Group. The operational insertion in impACT platform and development was operated by Group Data Office and CIB AI Platforms The tool is the result of internal development, using Open AI’s Chat GPT 4 o mini model via Azure API.
What role is left for the employee in the verification process to ensure the accuracy and reliability of the data entered?
The action of the RM is key in the process. Just like stated in its name, the AI Assistant is an assistant. It helps the RM reducing the time of completion of an ESG Assessment, but it does not replace them. They are the ones uploading the clients’ documents available first. Also, if the AI is providing answer suggestions to the RM, the RM is the one responsible for the final answer provided. The answers are not pre-filled, the RM must actively select to use the AI suggestion if deemed relevant, and they can edit the content generated by the AI. The RM can use AI to get answer proposals and can use the links to the passages of the documents giving the information translated by the AI as a suggestion, but they must double check the information provided.
What are the main key performance indicators (KPIs) used to evaluate the effectiveness of the ESG Assessment, and after several months of production, what are your initial results?
The AI Assistant has been used by 600 users since the Go live, helping them to complete almost 1600 ESG Assessments. The AI Assistant is used to complete around 35% of all ESG Assessments, and this rate is increasing gradually. We track the AI relevancy through quantitative and qualitative monitoring. And we also monitor the cost of running the LLM model in Azure, which is very limited in this use case. Overall, user adoption is good for this tool and user feedbacks are mainly positive.
What are the next steps or development paths?
We aim at continuously improve both the user experience (translation of AI suggestions in the languages used in the CPBS entities) and the AI performance. For the latter point, data scientists are evaluating the way to enhance the prompts to get even more precise suggestions from the AI.
