| Wavestone

At a glance

Sector: Non-profit

The challenge: Turning thousands of pages of ESG reporting into actionable insights.

The solution: Developing an AI-powered platform to analyze ESG data at scale.

Key results:

Successfully tested on water-related ESG data, analyzing tens of thousands of pages of reporting.
Standardized company data and cross-sector analyses generated automatically in seconds.
A scalable approach that can be extended to climate, biodiversity, circular economy, and other ESG topics.

Cross interview of Cécile Denormandie and Cédric Baecher

How can AI transform thousands of pages of ESG reporting into actionable insights? This is the ambition behind “Blue Wave by Wavestone”, an ESG analytics platform jointly developed by EPE (Entreprises pour l’Environnement) and Wavestone. Built on a multi-agent AI architecture, it can analyze thousands of pages of corporate documents and extract actionable insights directly.

Cécile Denormandie, Managing Director of EPE, and Cédric Baecher, Sustainability Partner at Wavestone, look back at the origins of the project, the approach implemented, and the lessons learned from this first experiment.

The challenge

Making sense of growing ESG data

How did this project come about?

CD: Companies today publish a huge amount of ESG data but that information is scattered across hundreds of reports thar are often too long, inconsistent and difficult to analyze at scale. As part of its 2026 work on “Water and Business Continuity,” EPE aimed to better understand and highlight its members’ practices in order to identify the major trends shaping their strategies for addressing water-related challenges.

Water is a strategic issue that cuts across many sectors and is high on the international agenda, especially in the run-up to the United Nations Water Conference in December 2026. It is also an area where a better understanding of risks, dependencies, and good practices can directly help companies strengthen their resilience as pressure on water resources continues to grow.

What made you see Artificial Intelligence as a relevant mean to address this challenge?

CD: With the acceleration of AI adoption, and given the many opportunities and challenges it brings, we wanted to explore how artificial intelligence could not only accelerate the processing of ESG data, but also help our organization turn a growing volume of information into actionable insights that support decision-making and action. This exercise enabled the EPE team to analyse tens of thousands of pages of data and generate insights on water-related issues at a scale that would otherwise have been impossible. The objective is not to replace human expertise, but to enhance it, freeing up CSR teams to focus on interpreting results, sharing best practices and creating value.

The solution

Co-creating an AI-powered tool

How did you go about developing the tool?

CB: From the outset, we chose a truly collaborative approach, bringing together EPE, Wavestone’s Sustainability teams, and the Wavestone AI Lab. The combination of EPE members’ hands-on experience of sustainability challenges with Wavestone’s dual expertise in technology and ESG was key to the project.

Wavestone also chose to support the experiment on a pro bono basis, as we saw it as a particularly interesting opportunity to explore how AI can contribute to addressing environmental challenges.

How did you turn that collaborative approach into a practical methodology?

CB: We started by defining a robust methodological framework. The thematic scope was aligned with CSRD requirements, specifically ESRS E3 on water and marine resources, and structured around ten main categories of analysis. We also defined the scope of documents to be analyzed and an interpretation framework to ensure the quality and comparability of the results.

The tool was then developed over two months, using a test-and-learn approach with multiple iterations between business and technology experts. This allowed us to progressively improve the quality of the analysis and the reliability of the results.

How does the tool actually work?

CB: The tool is built on an AI architecture made up of several specialized agents that work together to select the most relevant analysis methods, explore large volumes of documents, and generate answers that are systematically backed by sources. We paid particular attention to the robustness of the results by combining keyword searches, semantic searches, and in-depth analysis, together with regular validation by our business experts.

We also chose to prioritize data quality and comparability by initially focusing on CSRD-compliant sustainability reports and Universal Registration Documents, or URDs, whose structure makes comparative analysis easier. We then gradually expanded the scope to include other CSR and ESG reports published on a voluntary basis.

The results

AI alone is not enough to unlock sustainability data

What results have been achieved and what are the lessons learnt?

CD: In just a few months, the tool has processed a corpus of several tens of thousands of pages of ESG reporting documents from EPE members, extracting structured, comparable, and directly actionable insights. The platform now automatically generates standardized company profiles and can perform cross-sector or sector-specific analyses in just a few minutes.

It gives us a clear view of companies’ water strategies, enabling us in particular to assess their approaches to water, analyse how they address the risks and opportunities associated with this resource, and compare practices across different sectors.

CB: The key lesson from the project is that technology alone is not enough. The quality of the framework, data structuring, and business expertise remain essential to ensure relevant and reliable analyses.This experiment shows that a new generation of AI tools can help take sustainability data analysis to the next level.

As the volume of available information continues to grow rapidly, these approaches will make it possible to accelerate the production of insights that support decision-making, transformation management, and the sharing of best practices.

What are the next steps?

CD: We have already presented the tool to EPE members, who have shown strong overall interest in the development of AI-powered tools. We plan to present the approach and initial results at the DEFi4 2026 conference, during a workshop dedicated to “Water and Business Continuity.” We also aim to raise broader awareness of this initiative at the United Nations Water Conference next December.

In the longer term, we would like to extend the approach to other ESG topics, such as climate, biodiversity and the circular economy, and more broadly to any context requiring the analysis of large volumes of complex documents.

Companies today generate vast amounts of ESG data.

The challenge is no longer so much about accessing information, but about being able to analyse and compare it, and draw actionable insights from it.

Cécile Denormandie, Managing Director, EPE

Innovation lies not only in AI’s ability to analyze thousands of pages in a matter of moments.

Above all, it lies in its ability to turn this wealth of information into actionable insights that inform decision-making, by combining artificial intelligence with human expertise.

Cédric Baecher, Partner Sustainability, Wavestone