Net Zero Solutions: Using AI for Assessing Climate Transition Plans

In the first article of our series on climate transition plans, we focus on the topic of AI and how it can accelerate and improve the assessment of Net Zero alignment

For many investors, climate risk management means looking ahead and understanding companies’ preparedness to cope with future risks related to extreme weather and the transition to a lower carbon economy. Similarly, net zero investment strategies need to be informed by companies’ expected future emissions pathways. In both cases, companies’ disclosure of detailed climate transition plans (or a lack thereof) can be useful for informed decision making. 

Climate transition plans can expand an outsider’s view into the corporate strategic management of climate-related risks and opportunities. They can provide important insights into the credibility of carbon emissions reduction targets, including those reported to the Science-Based Targets initiative (SBTi). 

However, there are different perspectives regarding the characteristics of climate transition plans. Some of those organizations issuing reporting guidelines for companies see them as only one aspect of a company’s overall strategy. Other definitions focus more on the transformational idea behind them of shifting entire business models and aligning to 1.5°C science-based emissions pathways.

Solving a puzzle with many pieces

Based on different interpretations, a jungle of transition plan disclosure frameworks has emerged, with organizations including supranational bodies, industry groups, and NGOs issuing guidelines and recommendations with different focuses and varying degrees of detail. In Europe, regulatory traction is evident through the ESRS¹ and the CSDDD², while the FCA in the UK also plans to make transition plan disclosure mandatory.

At first sight, this complexity presents a challenge for many companies and investors. Fortunately, there are several areas of convergence regarding what represents a good transition plan disclosure by companies. For example, there is a consensus that disclosing emission targets, decarbonization actions, the use of carbon credits or negative emissions technologies, lobbying activities and internal climate governance is crucial for stakeholders.  

However, as highlighted in a recent study: “The climate transition is therefore a puzzle made of many pieces […] and each piece requires specific attention³”. This complexity of content makes analyzing transition plans a significant challenge for investors, especially when the investment universe extends beyond a few high-impact sectors where data is often more readily available. Leveraging technology, especially generative AI, can assist in this regard, streamlining sustainability data collection and interpretation while enhancing understanding of corporate transition strategies.

The AI-supported analysis of climate transition plans

For this purpose, Large Language Models (LLMs) can be utilized to extract and assess information from company CSR or other relevant reports. LLMs, sophisticated artificial intelligence models, are specifically designed to understand and process human language in a manner similar to how humans do. 

At Clarity AI, we believe LLMs hold the potential to revolutionize the automation of extracting, analyzing, and synthesizing transition plan information disclosed by companies. Therefore, we are using the technology to enhance our assessment of companies’ Net Zero alignment, according to the Net Zero Investment Framework (NZIF) by IIGCC. 

To achieve this, we have trained a model to extract and evaluate relevant elements from sustainability reports using the CA100+ Net Zero Company Benchmark, a framework recommended by the IIGCC⁴. This framework provides us with the rules by which our model evaluates the different elements of credible climate transition plans. 

Unlike other transition plan disclosure frameworks, CA100+ serves as an assessment benchmark, offering a more focused and prescriptive approach towards key details of transition plans. As such, we believe it can enhance precision in identifying the most relevant information and evaluating it. 

However, the application of AI for the evaluation of transition plans requires human assistance due to the high complexity and intricacy of the information. Especially because it involves multifaceted individual topics, our sustainability experts delve into them in detail and review numerous company disclosures to effectively train and continuously adjust the model, in collaboration with data scientists. 

The data obtained in this way allow for deeper insights into companies’ climate transition plans, even across large investment portfolios. For example, corporate decarbonization strategies can be analyzed more precisely based on individual core areas, such as the use of renewable energy and carbon credits, the expansion of green product lines, or a company’s engagement with suppliers and customers.

We will gradually integrate the data generated by AI on transition plans into our Net Zero portfolio analytics solution, empowering users to better identify and analyze leaders and laggards in their portfolios. It is important for us to ensure transparency regarding the statements made about the quality of the transition plans and the sources of information used. Our ultimate goal is to promote dialogue between investors and companies on climate risks and necessary adaptations, while supporting real economy transformations to achieve Net Zero alignment.

¹European Sustainability Reporting Standards (Disclosure Requirement E1-1). 

²Corporate Sustainability Due Diligence Directive (Article 15)

³ReclaimFinance (2024): Corporate Climate Transition Plans – What to look for

⁴Institutional Investors Group on Climate Change