ESRS and Gen AI: An exercise in trust and transparency

By Elisabeth Goos, Market Leader Sustainability Services EMEA & DACH, IBM Consulting

With COP28 just a few weeks away, business leaders have their eyes firmly placed on Dubai, preparing for the global summit and anticipating what announcements will be made.  

And as corporate involvement in the event has grown in recent years, so has the need for businesses to move away from meaningless greenwashing claims and demonstrate true, tangible progress. 

This scepticism is not unfounded. According to an IBM ESG Data Study, 95% of business’ have a sustainability strategy and ESG (environmental, social, and governance) value proposition in place, but only 10% said they have made real progress on it.  

It is therefore little wonder that only 20% of customers trust the statements companies make around sustainability. It poses the question: How can we trust that corporate commitments are translating into real action?  

Choosing the right tool for your objective 

A wave of new sustainability regulations have been announced to address this very issue. The European Sustainability Reporting Standards (ESRS), following the Corporate Sustainability Reporting Directive (CSRD), represent some of the most far-reaching reporting requirements so far. They call for a comprehensive assessment of business impact across multiple environmental and social dimensions, including climate, water, biodiversity, water and circular economy. This comes with detailed requirements on data collection, verification and performance management.

These regulations offer a real opportunity for businesses to prove their sustainability progress in a comparable way. It is also a chance to modernise and harness innovative technologies in the process. 

One technology that holds great promise in the area is generative artificial intelligence (Gen AI). 

Gen AI models – referred to as foundation models – can be trained on huge volumes of data to learn the patterns and structures of language, then fine-tuned on data from specific tasks or domains to generate output on certain contexts or objectives. 

Thanks to their versatility, these models are being heralded as a key tool for businesses looking to successfully complete ESG reporting and gain deeper insights into their own data.

The business benefits of reporting

The CSRD, with the standards published by the ESRS this October, requires any EU company with over 250 employees, more than 40 MIO annual revenue and 20 MIO in total assets, to file an annual report using the final ESRS guidelines, including disclosure of how sustainability influences their business as well as the company’s impact on people and the environment. It calls for more granular information, operationalised KPIs and asks for details in many areas not (yet) covered by other frameworks (e.g. Biodiversity). 

The standards strengthen the requirements around reporting, establish legal synergies with upcoming regulations on green-washing and supply chains, providing the opportunity for best performers to demonstrate their credential, as the standards ensure transparency and full comparability across companies. They help investors and other stakeholders receive information, backed by data, that they need to assess environmental and social impact, financial risk around climate change, and sustainability opportunities within a business.  

What’s more, the new frameworks will bring a wealth of audited, comparable, trustworthy data that could energise the whole sustainability discussion and re-introduce trust back to corporate sustainability.

It’s all down to data

While the benefits are clear, the standards do pose a challenge. 

Data has already proven to be an ongoing issue for businesses, with 41% of executives believing that inadequate data is holding back their ESG progress. Given ESRS requires new data sets on new issues, such as pollution and biodiversity, this pre-existing issue will be exacerbated. 

The complexities associated through quantifying this data also make it intractable using traditional reporting practices. Manually sifting through data to calculate carbon emissions and biodiversity, for example, is not only error prone, but unwieldy, time-consuming, expensive, and can result in an unreliable audit trail.

Implementing a specialised platform – backed by AI – to capture data and calculate emissions, monitor sustainability initiatives, and assess supply chain feedback makes the process easier, more reliable and transparent. By harnessing the right technologies, businesses can create value, not only in the exercise of reporting, but in the insights and recommendations that come after.  

The role of generative AI

The foundation models of Gen AI are powerful tools for sustainability reporting for two main reasons.  

The first lies in the ability to rapidly analyse patterns and structures of language, as well as numerical data. Businesses will be required to report on brand new areas – such as pollution – and existing data will likely be varied and disparate. Yet Gen AI can pull this existing data from historical reports, documents and databases, that was not previously possible. 

One example of this strategy can be seen in a project from natural oil and gas company Wintershall Dea. The organisation had a vast subsurface knowledge base containing hundreds of thousands of documents, which was inaccessible to traditional AI and advanced analytics. However, through a generative AI -based knowledge extraction tool from IBM Consulting, the business was able to ask specific questions of the provided knowledge base (tables & text) and evaluate whether certain sites are suitable for carbon capture storage projects in terms of profitability, effectiveness, and safety. 

This application of Gen AI can be harnessed by businesses looking to extract data and insights from documents and reports that traditional AI cannot utilise. 

A second benefit is Gen AI’s unique capability to recognise patterns and quickly generate insights and recommendations for performance improvements. This is invaluable for the ESRS. Although the directive only requires the initial reporting, Gen AI will help businesses build upon this, drive insights, and create actionable frameworks that will move the needle on sustainability progress now and for future reporting requests. 

The foundations of trust and governance 

Sustainability reporting and Gen AI go hand-in-hand. They represent two of the biggest priorities for European CEOs in 2024 and both must have trust, transparency, and governance at their core.  

To make a success from the ESRS, businesses need to focus on the governance of the reporting, setting up groups who oversee the processes and creating the foundation to translate sustainability business strategy into an investable framework.  Likewise, a successful AI strategy has governance at its core – based on transparency, human oversight, security and trust. 

What is clear is that the companies that fare best will be those that embrace AI to ensure their business benefits from the latest wave of regulations and is ready for the sustainable economy of tomorrow.

Related Posts
Others have also viewed

Meet the trailblazing women collaborating to save the ocean and increase gender diversity in STEM

In Mauritius, Scotland, Manchester, London, and Australia a group of award-winning women scientists and experts ...

STUDY: UK transport and logistics industry faces sustainability gap admist AI adoption

HERE Technologies, the leading location data and technology platform, today unveiled insights from its latest ...

Einride, Mars partner for Europe’s biggest road freight electrification in FMCG industry

Einride, a freight mobility company that provides digital, electric and autonomous technology, has partnered with ...

BCG and Climeworks sign historic 15-Year carbon removal agreement

Climeworks, a global leader in carbon removal via direct air capture technology, and Boston Consulting ...