More than speed: rebuilding trust in ESG Data
As skepticism around sustainable investing grows, Borja Cadenato, Clarity AI’s head of AI and Data Solutions tells Environmental Finance how AI could help rebuild trust by grounding it in data
Environmental Finance: How can AI help rebuild trust in ESG Data?
Borja Cadenato: ESG has faced significant skepticism, partly due to greenwashing and vague marketing claims, undermining investor and consumer confidence. However, the core principles of ESG – and sustainability more broadly – remain crucial.
There’s an opportunity to address this challenge through AI. Both AI and sustainable investing are at defining moments: AI is already transforming every industry with greater speed, scale, and insight, while sustainable investing faces a credibility test amid growing ESG backlash.
This is a chance to build back better – not by replacing human judgment with AI, but by enhancing it. AI enables us to analyse vastly more data, spot patterns we might miss, and reduce unconscious biases to uncover a company’s true impact. It empowers people to make better-informed decisions – grounded in facts and guided by today’s context: sustainability.
EF: How can your GenAI offering enable investors to interrogate, not just consume, data?
BC: Traditionally, investors have relied on aggregated ESG ratings to guide investment decisions, assuming one-size-fits-all analysis would be enough. However, there’s a growing shift away from these 'black-box' ratings as investors seek direct access to the raw data behind them. In fact, 70% of investors have expressed concerns that relying solely on ESG ratings could encourage companies to prioritise reputation management over creating real, measurable impact, which risks turning ESG into a 'check-box' exercise.
Through GenAI, analysts can engage directly with the data. Our AI Assistant helps them gain a deeper understanding, test hypotheses, refine their assumptions, and perform customised analysis. The result is akin to having a highly skilled expert by their side, guiding them through the complexities of sustainability analysis.
An example is evaluating climate transition plans to understand how a company plans to reduce emissions, aligns with regulatory expectations, and meets net-zero targets.
Sustainability expertise is essential in assessing whether a company’s transition goals are ambitious enough and if its actions will meet those goals. According to our research, only 40% of high-emitting companies disclose and quantify the impacts of their decarbonisation actions, revealing a significant gap in transparent information, undermining investors’ ability to assess a company’s true climate ambition.
To address this, we developed a framework that assesses the credibility of companies’ transition plans and their likelihood of success. Grounded in international standards like the Transition Plan Taskforce and Climate Action 100+, this framework processes answers through Large Language Models (LLMs), applying sector expertise and scenario data.
The resulting insights are presented in detailed company briefs, which provide a credibility score, supporting data, and AI-generated summaries highlighting evidence from the company’s disclosures. This offers investors a comprehensive, 360-degree view of a company’s transition plan, allowing them to make more informed, data-driven decisions.
EF: What can enhance integrity in this space?
BC: Data maturity is a key factor in restoring trust. In the past, ESG data was often sparse, inconsistent, and hard to compare across companies. Reflecting this, over 70% of institutional investors have cited inconsistent and incomplete ESG data as the primary barrier to effective sustainable investing. AI can address this by processing a broader range of data sources, including news, social media, and independent reports to provide a more complete and balanced view of a company’s sustainability profile.
EF: What should responsible AI use in ESG data look like?
BC: To mitigate risks like bias or hallucinations, we combine agile AI development with Subject Matter Experts (SMEs) who validate outputs, provide feedback and guide improvements. SMEs also diagnose the problem and define the objectives, ensuring we apply AI to the right challenges – not just for the sake of it. This approach has led to a 40% increase in accuracy compared to other models like ChatGPT.
We must also be mindful of the environmental impact. By being intentional about when and how we use AI, we can reduce unnecessary processing and align technology use with our sustainability goals.
For more information, see: https://clarity.ai
