This report is our inaugural whitepaper where we examine how artificial intelligence (AI) could be used in ESG (Environmental, Social and Governance) evaluation to aide asset managers' efforts in assessing corporate behaviors along the ESG spectrum.
Traditionally, ESG scores are provided by vendors such as MSCI and Sustainalytics, where the core source of information used is company annual disclosure, which is voluntary. With the rise of AI applications in many industries, it makes previously impossible tasks now possible thanks to the scalability and speed that AI brings. Specifically in ESG, where objectivity and transparency are often the key asks from investors, this is where the AI-led ESG providers can fill the gap as they collect and process relevant ESG data in a coherent and unified manner. Their platforms offer the granularity and quality of information that investors can interpret and integrate into their investment processes, providing more up-to-date ESG assessment on companies.
In this report, we have selected three AI-led ESG data providers who have distinctly different methodologies and constructions of their own ESG scores and ratings. We examine each vendor in turn and provide details on their sources of information and the application of AI / Machine Learning techniques in their process to convert massive amounts of data into final ESG scores. In addition, we look at the coverage, tilts and strategy performance based on the ESG scores from each vendor.
The purpose of the report is to shed some light on how each provider processes the large amount of alternative data sources available and transforms it into usable insights. While we do not think that AI-led ESG evaluations will completely replace disclosure-based ones (given AI relies heavily on media and news sources etc) we believe it could be a powerful tool to combine human analysis for longer term company performance with AI analysis for shorter-term evaluations where more exhaustive ESG assessments could be achieved.