How can generative AI help with your corporate-sustainability efforts?

Generative AI can play a strong supporting role in helping companies meet these CSRD requirements. In addition to reporting quantitative data, companies must submit qualitative analyses of their ESG performance. Through their ability to quickly analyse text, Large Language Models (LLM) can provide both structure and content for this qualitative reporting.

“Many employees with responsibility for ESG are new to the field. The CSRD guidelines can be overwhelming. Generative AI is very good at analysing this kind of content and summarising the main points, especially if your prompts are good. The technology can even support with initial drafts,” says Merwin Olthof, Sustainability Lead at Fellowmind.

Unlike other business functions that may have decades of data in various formats, sustainability reporting allows companies to start afresh. There is an opportunity to shape the practice based on the common language mandated by the CSRD.

“With companies publicly disclosing their climate-risk strategy and opportunities, for example, there is a huge pile of relevant sustainability information being built up. This will become a valuable source for the AI, with bots increasingly adapted to specialised use cases such as benchmarking your efforts to competitors,” explains Olthof.

Sustainability-focused AI tools emerge

ESG-specific AI is already available through Microsoft Copilot for Sustainability. It’s an AI-powered tool integrated into Microsoft Sustainability Manager, the cloud-based solution that helps organisations to manage and reduce their environmental impact.

Microsoft Copilot for Sustainability enables the use of natural language to set up sustainability tracking and reporting systems, as well as to analyse ESG documents. Users can interrogate the data, create calculation models, generate emissions reports and more.

“Copilot for Sustainability is a valuable resource for companies across all industries, as it can greatly democratise the use of digital tools and information. This is especially needed in the agricultural sector, where the adoption of digital tools has been quite slow. Now LLMs are lowering the barrier for farmers to access digital information,” says Olthof.

“The agri-food industry is very much interested in AI. The sector has some strict emission-reduction targets, so there is a lot of work to do,” he says. “In the future we can envision a scenario where farmers can ask the AI questions about different variables – such as soil properties and climate – so that Generative AI can assist with risk assessment and meeting carbon-reduction targets.”

Like other sector-specific AI implementations, the outcome depends on the quality of the underlying data. As more CSRD content is generated and companies harmonise their data structures accordingly, AI-based sustainability tools will become more reliable and useful. The first CSRD reports will start to come out in 2025 – as part of companies’ 2024 financials –  so generative AI will be able to quickly learn the required formats and support others who need to report in the correct way.

“This initial wave of companies doing their CSRD reports are really performing a lot of the heavy lifting. Risks have to be described in a certain way and careful explanations are needed as to why a company is reporting on a specific topic. Opportunities need to be explained properly too. All this represents a huge amount of input for specialised generative-AI sustainability models to learn from,” says Olthof.

AI-driven support for suppliers and on the factory floor

The potential with AI becomes even greater reporting on Scope 3 emissions – i.e. the environmental impact of a company’s supply chain. Getting emissions data from third parties can be a huge task, especially in industries with diverse supplier networks that stretch across the globe.

“Scope 3 data availability is a big concern. Large retailers, for example, are supported by hundreds or even thousands of suppliers. These companies will need a lot of support in gathering third-party data,” notes Olthof.

AI-based chatbots specialised in sustainability can assist suppliers in providing the required information. Bots can be trained to ask the right questions, as well as to guide suppliers towards uploading the necessary information. AI can also quickly read policy documents from suppliers and transform the information into CSRD-ready inputs.

In the end it’s the employees of a company – particularly those involved in production – who are responsible for implementing sustainability measures. Another way in which generative AI can help is through the implementation of chatbots that answer workers’ questions on different sustainability topics. This would negate the need for companies to create complex instruction manuals in multiple languages.

Do you need help with your sustainability strategy?

Whatever industry you work in, Fellowmind can help to harness the potential of AI to meet your company’s sustainability goals. Our deep expertise in Microsoft platforms and tools empowers your ESG team with enhanced data gathering and reporting across the entire supply chain.