Indirect influence - Direct Climate effect
Purpose and goal
The aim of the project was to develop a prototype of an AI solution to identify how well companies report on climate issues. More specifically, the project has developed an AI solution that focuses on how companies work with climate-related risks and opportunities, based on reporting in line with TCFD´s recommendations. By categorizing available data, automating analysis using AI, investors have the opportunity to influence companies in a more sustainable direction. We have succeeded in developing a beta version of a service that can be scaled up and used by investors globally.
Expected results and effects
If investors have easy access to comparable data on how companies perform based on the climate issue, their opportunities to influence them in a more climate-smart direction will increase. The expected effects we hoped for with the project were that that information can be made available and comparable. The results from the project have given us knowledge to scale up the AI solution, so that investors can use the developed ML algorithm on annual reports to get a comparable data with a focus on companies´ governance, strategy, risk management and targets and metrics in climate.
Planned approach and implementation
This project included four work components 1) Scraping; Collection of relevant data (extract text from pdf) 2) Categorization; Data is categorized and analyzed - for the algorithm 3) The algorithm is trained- deep learning 4) Analysis of a selection of companies from the AP7 list - results are presented. During the project, we had to compare and analyze several different AI services and find the right skills to assist in the work with the knowledge, drive and close cooperation required - this meant that we had to extend the project with a shifting schedule and focus as outcome.