ESG has emerged as a key focus for businesses worldwide. Presently, all major EU corporations are in the process of applying the Corporate Sustainability Reporting Directive (CSRD) to their 2025 reports. While the CSRD primarily applies to EU-based companies with an annual turnover exceeding €150 million, it serves as a precursor to broader sustainability regulations. For example, 16 pieces of legislation affecting the entire value chain of the retail industry, from product conception to marketing, are due this year.
This transition is already influencing various industries. In addition to regulatory requirements, investors are increasingly using ESG ratings to guide their investment decisions, and this behaviour will only increase as ratings are standardised. Furthermore, customers are growing more conscious of ‘greenwashing’ practices, which will potentially alter their purchasing and investment decisions based on how companies are rated.
To prepare for these transformative shifts, businesses must elevate their ESG protocols. This isn’t only important for meeting regulatory requirements but is also fundamental for a business’s comprehensive auditing and strategic plans. Enhanced ESG planning enables finance departments to integrate ESG considerations into financial frameworks, gain deeper insights into operational efficiency opportunities and costs, and inform strategic decision-making processes. Improved margins and cash flows, streamlined reporting, and greater visibility into a business’s performance are the result.
Establishing the link between a company’s financial plans and its ESG framework has been a challenge for FP&A. To address this, businesses need to support a solid strategy with modern and innovative tools. Traditional data analysis tools and spreadsheets are simply not sufficient.
McKinsey’s research underscores how cloud-powered technologies like artificial intelligence, machine learning, and the Internet of Things can accelerate almost half of businesses’ net zero initiatives (47%). This highlights the critical role of technology in helping organisations meet sustainability and regulatory requirements while competing effectively.
AI for data processing
Data surrounding ESG frameworks is complex, as it is derived from multiple sources and comprised of both structured and unstructured data created across an entire business value chain. For example, the environmental component of ESG must consider each step in a supply chain, from shipping and packaging to cloud usage and employee commuting.
To tackle the challenges associated with deriving value from ESG data, AI techniques such as Generative AI can be applied. AI is well suited to automating previously manual efforts in the gathering, merging, and analysis of data. It can combine ESG data with other related data sources, summarise vast amounts of text, and generate actionable insights in a fraction of the time required by traditional methods.
However, AI-derived outcomes can only be as good as the data to which they are applied. To maximise AI’s benefits, companies must ensure access to clean, accurate data from both internal and external sources. Accelerating access to bad data through natural language is obviously detrimental to any AI initiative, and thus it is key for organisations to have a solid data strategy in place before any AI techniques are leveraged.
When used correctly, however, machine learning, generative AI, and even video/image processing can provide teams with a broader and more comprehensive understanding of their data. Organisations are already using classical AI techniques to process ESG data across the business, and while Generative AI use cases are emerging, initial results are promising as natural language prompts allow analysts to interrogate data and generate plain-English analyses, automate ESG report creation processes, and reduce errors.
This level of data integration and analysis is key for creating an overall ESG plan. Businesses must be able to track, monitor, and adjust according to varying sustainability goals. The best way to use AI is to collect, standardise, and aggregate data for each metric involved in ESG reporting. Once this is in place, businesses can more easily understand their individual ESG targets. An intelligent planning tool can facilitate these tasks, especially if it provides capabilities beyond simple “AI-washing” and allows businesses to truly work with AI in a seamless and integrated fashion.
But can we trust AI yet?
Despite its transformative potential, trust in AI remains a concern. While companies are under pressure to improve productivity and efficiency with techniques such as AI, some mistrust around the technology exists. This is evident in the industry, as in March, the European Parliament approved the Artificial Intelligence Act, which takes a risk-based approach to ensure companies release products that comply with the law before they are made available to the public.
ESG regulations demand that companies have audit points that are fully explainable and transparent, and creating matching processes should be a top priority for finance teams deploying AI. High profile examples of generative AI ‘hallucinating’ stem largely from its probabilistic behaviour along with the quality of the inputs used to pre-train, ground, prompt, or fine-tune the models.
ChatGPT, for example, is trained on data from the publicly available internet, which can contain inaccuracies and biases. To improve outcomes, organisations can look to use specialised large language models, where models have been tuned or trained with proprietary or domain-specific data. This can reduce hallucinations, again assuming that the data is of sufficient quality and accuracy, as specific context is readily available to the models.
Data teams can also employ techniques such as RAG, ReAct, model fine-tuning, or prompt engineering to improve the quality of their responses. And, as models can suffer from ‘AI drift’, whereby the model becomes less accurate over time, it’s important that humans are able to easily verify outputs via grounded references and other background sources provided in the model responses.
Companies can mitigate these flaws and optimise their AI use
Businesses have ample opportunities to safeguard against the potential pitfalls of AI, as carefully compiling high-quality datasets will set a strong foundation for any AI initiative. Customisations can be introduced to your AI integration points to improve overall accuracy, and solutions can be deployed with appropriate management and monitoring capabilities to ensure continuous, accurate outcomes.
Furthermore, clarity and consistency of standards within emerging ESG regulatory frameworks will be important to ensure deeper standardisation across datasets, reports, and metrics, and make it easier to compare information from numerous sources. Timelines for achieving these objectives, however, will depend on several factors, including the pace of technological advancements, industry adoption rates, and the formulation of regulatory guidelines.
The integration of AI into ESG functions holds the promise of resource savings for businesses. However, before embarking on this journey to streamline ESG reporting through technology, a thorough assessment of available data and an evaluation of how technology can foster new best practices are essential steps.
Nelson Petracek is Board’s Chief Technology and Product Officer. He has over 20 years of experience, including at TIBCO and Informatica, building world-class technology teams at high-growth organizations. Throughout his career, Nelson has delivered systems for the digital transformation age, drawing upon his deep knowledge of analytical applications, APIs, cloud, event-driven architectures, and other emerging technologies.