Can AI close the gaps for full compliance with SFDR?

by Andreas Pade, Transformation Expert and Project Manager in the Financial Industry 
published by IT Finanzmagazin  

From sustainability reporting under CSRD (Corporate Sustainable Reporting Directive), around 15,000 companies and their suppliers in Germany are affected alone. The disclosure of ESG data according to SFDR (Sustainable Finance Disclosure Regulation) is mandatory for financial market participants within the EU. Some software companies already offer solutions intended to simplify the reporting process – and AI is often a crucial part of these solutions. But in which areas is AI particularly useful in sustainability reporting? Providers such as Cleversoft, Code Gaia, IBM, LPA, UBRICH, SAP Fioneer ESG Solutions, and the Zeidler Group offer answers.  

Artificial Intelligence is increasingly seen as an interesting option for better reporting. Why? Advanced analytics, AI, and blockchain are changing the way companies approach ESG reporting. These technologies allow for the automation of data collection and analysis, reducing human error and increasing accuracy. Sophisticated tools provide real-time insights into ESG performance, enabling proactive decision-making and continuous improvement. Particularly, blockchain offers transparency and traceability, ensuring that every entry in the ESG report can be verified and traced back to its source. 

How software providers approach CSRD/SFDR reporting and AI 

Which software providers that currently offer solutions to support CSRD/SFDR reporting are already using AI, and for what purposes? What are the providers planning for the future? Representatives from various market participants indicate that AI is currently primarily used in the areas of data handling and text generation. 

Data handling 

Handling ESG data without AI support is very labor-intensive. Thus, the benefit of using AI is significant. The constraints here are driven more by regulation than of technical nature (e.g., audit obligations). 

• According to Markus Breuer (COO) of the ESG Data Exchange Network provider UBIRCH, their AI-powered ESG Assistant is trained to answer specialist questions (primarily regarding GHG emissions) as part of a consulting function in creating an SFDR report. This allows suppliers, particularly those who provide data to customers that are already under reporting obligation, to do that much easier and significantly reduce the scope of the required consulting. 
• According to Maria Patschke, CEO (SAP Fioneer ESG Solutions), AI is used in SAP Fioneer’s ESG solution to detect inconsistencies in input data and calculated results. Furthermore, AI supports quality assurance for determining financed GHG emissions, as AI also makes suggestions for quality improvement. Due to the regulatory requirement of auditability, AI is currently not explicitly used for data calculation at SAP Fioneer. 
• In the case of the sustainability-as-a-service provider Code Gaia and the IBM Envizi ESG Suite, the AI reads financial data from the accounting software and categorizes it with respect to the correct emission factors for Scope 1-3 (IBM only Scope 1) according to the GHG protocol based on the activities and quantities carried out. Approval of the values by the user is required for the IBM ENVIZI ESG Suite. 
• According to Markus Diesing (CPO), RegTech provider LPA is currently working on a solution in which AI supports and accelerates the mapping of customer data to the LPA input format. 

Text generation 

One of the strengths of AI is generating texts based on data. 

• As Markus Adler (Co-Founder) reported, Code Gaia’s AI provides drafts/sample texts for reporting that customers can import into various disclosures and then edit. Users can add additional notes with which AI refines the text drafts to make them more specific. The content created by the AI must be accepted/reviewed by the users. 
• SAP Fioneer will use Natural Language Processing (NLP) in the future to generate the narratives of the ESG reports.  
• Similarly, the RegTech provider Cleversoft, according to Florian Clever (CEO), will also use AI support in the future for creating narratives for SFDR reporting. 

What will the future of AI application in CSRD/SFDR reporting look like? 

• In the near future, UBIRCH sees the possibility of using AI to fully query the necessary data in dialogue instead of via a form. In the field of data collection, Code Gaia sees opportunities for AI, for instance, in reading web content, Robotic Process Automation, and advising on possible data sources. 
• UBIRCH and LPA are planning to use AI to check and validate the plausibility of the data supplied. Code Gaia envisions applying AI to quality assurance processes and anomaly detection to identify incorrect entries and results. 
• According to Oliver Gahr (Sustainability Lead, IBM Technology EMEA), AI is expected to increasingly help close data gaps. He anticipates that with AI, a streamlining of SFDR reporting could occur, focusing on the contents that must currently be reported. 
• In the future, Elisa Forletta-Fehrenberg (Head of ESG at the Zeidler Group) sees the possibility of using AI to reuse data and information belonging to SFDR reporting for reporting for example under CSRD.
• SAP Fioneer and Code Gaia consider individualized strategy proposals with recommendations for improving ESG scores in line with the requirements of the CSRD and SFDR to be feasible. 
• SAP Fioneer can also envisage the following developments in the next three years: Extended scenario analysis, in which different ESG future scenarios for regulatory changes can be simulated with the help of AI. AI-supported adaptation of reporting to regulatory changes or new requirements from the CSRD/ SFDR. 

Conclusion

We are clearly on a long journey in the area of the application of AI for CSRD/SFDR reporting, which has only just begun. 

Looking to the future of the use of AI in CSRD/SFDR reporting, further development will depend largely on the regulatory framework in place by then. Providers are pursuing numerous approaches to further expand the use of AI. Developments are planned in particular regarding data quality, text generation, sustainability strategy, improving the efficiency of data utilization and user-friendliness. The independent adaptation of the reporting system to regulatory changes is also on the agenda. Can CSRD/SFDR reporting be fully automated in the future with the help of AI? The technical and regulatory developments of the next few years will show this. 

About the Author Andreas Pade 

Andreas Pade is a transformation and product expert in the financial industry. Most recently, he was responsible for setting up the fund administration at CACEIS in Germany and the integration and strategic realignment of the regulatory reporting sector for asset managers at RegTech provider LPA. His focus topic is the transformation of the financial industry, with an emphasis on leadership, automation, digitization, AI, and ESG.