Use of AI in sustainability reporting - Part I: The report preparation process
ESG reporting requirements have been considerably curtailed within the scope of the EU Omnibus Regulation. Yet, producing a good sustainability report will still possibly involve a very significant amount of work. This can be again greatly reduced by using artificial intelligence (AI). In Part I of this report, we first discuss the sequence of steps in the report preparation process. In Part II, in next month’s PKF Magazine, a description will follow of how and which AI tools can be used as aids in the individual process steps.
Introduction
AI tools are software applications that are based on artificial intelligence (AI). These tools are designed to simulate human intelligence, to learn complex tasks, to analyse and to develop further autonomously. Well-known examples of such AI software are chatbots and generative AI for text creation and image editing. The best known is presumably the tool ChatGPT. This is a language model that has been trained to understand human language and to respond in natural language.
Within the scope of this report, we discuss how deploying AI tools can aid sustainability reporting. First of all, in section 2 below, we outline the sequence of steps in the process for preparing a sustainability report. In Part II, this will be followed by an overview of examples of AI tools for the individual steps. To conclude we will consider the limits of the use of AI in sustainability reporting.
The report preparation process
Overview
The preparation of a sustainability report is a complex process; it includes several phases that build upon one another and it is time-consuming. Once the general applicability of the reporting requirement has been clarified, three core phases in particular appear to be especially challenging:
- the materiality assessment,
- the collection and evaluation of data, and
- the actual preparation of the report.
The individual phases
1. Defining applicability
First of all, it will be necessary to check if a reporting requirement exists, or if there should be voluntary reporting. If this is the case, then a structured schedule as well as clearly assigned responsibilities will be needed.
2. Materiality assessment
Important stakeholders across the value chain have to be determined and an initial overview of the potential key issues (impacts, risks, opportunities - IROs) has to be created. Subsequently, an assessment is made of:
- the financial and non-financial effects,
- the possibility to correct these, and
- their probabilities of occurrence.
The results will determine the extent to which there will have to be reporting on ESG issues.
For companies, carrying out a materiality assessment involves major challenges. The process ranges from drawing up an initial list of issues (IRO long list) right through to specific analyses, such as climate scenarios, for example. A major difficulty consists in correctly selecting the relevant issues. Large companies often struggle to find the right level of detail. By contrast, small and medium-sized enterprises face the challenge of having to implement the stringent requirements of the standards with limited resources in terms of time and personnel as well as a lack of expertise.
3. Objectives and measures under the ESRS
The European Sustainability Reporting Standards (ESRS) are the requirements for non-financial reporting and, as such, can be compared with the IFRS for financial reporting. The ESRS require companies to define specific objectives and appropriate measures for material sustainability issues. The sustainability strategy should be embedded in the overall corporate strategy.
4. The collection and evaluation of data
The relevant data has to be gathered on the basis of the materiality assessment. This will include both qualitative as well as quantitative metrics that are appropriate for the ESRS requirements. In doing so, a cross-check against already existing company data can help to identify any gaps and to close these specifically. The most challenging aspect here is the variety of and complexity of data sources. Locations, departments, supply chains - all these areas supply data that needs to be combined in a standard format. Companies that operate internationally, in particular, will need to put in a considerable effort. However, the biggest obstacle is not the lack of data, but rather their structuring, transparency and usability. One example is the recording of all the emissions and converting them to CO2 equivalents.
5. The preparation of a sustainability report
The large amount of information needs to be transferred into a clear and easy-to-understand report. In doing so, contents from different sources must be combined, checked for consistency and communicated in comprehensible and correct language. However, the text creation is especially time-consuming because the formulation must be precise, comprehensible and free of inconsistencies.
6. Publication and integration into the business processes
Insofar as there is a sustainability reporting requirement, the report must form part of the management report and be published in the European Single Electronic Format (ESEF). After publication, it is crucial to continuously develop sustainability management further and to optimise the existing processes on an ongoing basis.
The need for support software
As sustainability reporting entails a considerable effort by companies and their employees there is a great demand for software solutions in order to address these challenges. Conventional tools can be used here to support or simplify the individual phases of the process. However, particularly promising software solutions, which are increasingly based on AI, are those that reduce manual tasks and make the processes more efficient for the users.
In Part II of this report, we will explain in greater detail how AI-assisted tools can effectively provide support for sustainability reporting. Here they can help to efficiently process large amounts of data, comprehensibly interpret the regulatory requirements and prepare standard-compliant reports. Automation facilitates process optimisation, resource saving and enhances the quality and the transparency of the reports.
To begin with, it should be noted that, in practice, the tools are rarely solely AI. In fact, the software solutions that are used integrate specific AI functions into their applications. These solutions support the entire reporting process - from the materiality assessment, including the collection and evaluation of the data, through to the actual preparation of the report - and use AI to make the individual sequences of process steps more efficient and precise.
