ERM produces ESG assessments within two business days using Azure Machine Learning

ERM leads the way in ESG scoring with machine learning

As institutions look toward responsible investing to support social causes and promote sustainable business practices, environmental, social, and governance (ESG) reports and scores have become a key tool to screen the value and integrity of potential investments. For over 50 years, ERM has paved its way as a global leader in sustainability consulting and now offers ESG Fusion as a digitally enabled solution to complement existing consulting services.

 ERM developed its scoring primarily for privately-owned companies, where ESG data is scarce, and there is a need for an efficient tool that can provide quick analysis for investors. Using ESG Fusion, ERM helps investors to understand the key ESG risks for informed decision-making through the screening stage and in support of the due diligence process. This enables investors to take ESG risks and opportunities into critical consideration during the investment decision period.

ESG Fusion determines a rating, evaluating companies on their performance against key indicators in critical areas like the environment, climate change, human rights, and DEI (diversity, equity, and inclusion). Investors can look at this ESG score and decide whether an investment aligns with their values and is worth backing based on its overall risk. Similarly, a company can obtain its own ESG score to understand how its business practices may be out of alignment with global trends and requirements as well as their own company values and goals—and how this might affect future partnerships. 

Finding and making sense of all this ESG data is a considerable challenge. Accessible data in the public domain can vary in volume and type depending on the industries and companies. For researchers, manually sorting through enormous amounts of data and data sources to assemble a full picture is a near-impossible task.

Accelerating ESG Fusion to automate and scale audits

ERM utilizes the organizational expertise of ESG Fusion to save time and provide the most accurate data available. This is where ERM believes it has an edge on its competition. "ERM consultants are fully aware of competitors in the market who are providing ESG data,” says Gareth Lane, Business Unit Managing Partner at ERM, “but they're always coming back to ESG Fusion because it is the most comprehensive report that they feel is available to them in the market.” 

ESG Fusion crawls through search engines, corporate documents, sustainability reports, regional websites, adverse media news, and any potentially valid information in the public domain that sheds light on a company’s ESG practices or lack thereof. When it comes to private companies, ERM clients can provide documents and borrower questionnaires to the Fusion research team to include information that is not publicly available during scoring. 

ESG Fusion filters, sorts, and analyzes data from a centralized Microsoft Azure database. Azure Machine Learning both powers the analytics process and gives developers the freedom to bring their own natural language processing (NLP) and AI-driven data curation tools into the fold. The ESG Fusion and Azure Machine Learning analysis moves through three processes. 

To begin with, the system looks at the industry classification of the assessed company to identify any industry-inherent risks. It then focuses on company-specific risks, more specifically on incidents or accidents which can help illuminate how a company’s operational risks could lead to real-world consequences. The third process assesses management maturity, which examines how a company manages risk, responds to events, and proactively creates positive impacts through its products and services. In total, ESG Fusion considers as many as 100 indicators, which include greenhouse gas emissions, water emissions, transportation risk, and supply chain risk, and more.

“Using Azure Machine Learning, ERM built ESG Fusion with responsible AI practices,” Lane. “The system crawls the web and third-party data looking for up-to-date articles, websites, and coverage regarding the assessed company’s sustainability efforts based on robust and transparent methodology.”

Gaining meaningful analysis

Together these data points offer a historical understanding of a company while showing the growth of its ESG efforts over time. ESG Fusion helps in prioritizing diverse sources that could be relevant to evaluating a particular ESG topic. Human expertise then verifies the evidence that applies to the relevant ESG topic. Holding AI accountable to human expertise is a core principle of responsible AI, uncovering technological biases and supplementing the expertise of veteran specialists rather than replacing them.

“All of this is developed with subject matter experts—who have the real knowledge auditing these companies—and they have decades of combined experience,” says Pooja Nirwal, ESG Fusion Senior Consultant and Subject Matter Expert at ERM.

ESG Fusion presents critical sustainability information to ERM experts for review before it shares final ESG evaluations with potential investors. Following a review, clients receive a detailed report on the company investigated. This innovation enables analysis at an extraordinary scale, as ERM is now able to find data on more topics and evaluate them in less time.

“Before ESG Fusion, consultants were independently looking in their own areas,” says Lane. “They would have guidance from the partner in charge, but it was time consuming. This is where we saw the standardization of technology that could provide a vehicle for us to take these reports to the market. And that vehicle is developed with Microsoft Azure Machine Learning.”

Nirwal concurs. “Data is at the heart of the tool, and it is amazing to see how we have created the universe of data sources using ERM`s expertise for private companies over the years, to provide meaningful analysis to them.”

Driving ESG data to the next level using Azure

ESG Fusion is powered by the Azure technology stack, with the front end (Single Page Application framework) and back end (.Net framework) hosted on Azure App Service, the machine learning processing engine scaled on Azure Kubernetes Service and communications with external services managed with the Azure Service Bus and API management tools.

When ESG Fusion debuted on the Azure Marketplace, it came as a pleasant surprise that its technology was adopted so quickly. “Success has been quick,” says Lane. “This has put us way ahead of schedule for this financial year and we are now managing the level of reports that we were hoping to achieve a lot quicker than we initially anticipated.”

As the ESG Fusion data universe continues to grow, it will supply even deeper insights into business and industry practices. Not only are ERM consultants gaining a valuable tool right out of the gate—clients can measure company performance over time and across peers.

“Microsoft and ERM have really demonstrated the talent, skillset, and expertise that goes into building this technology,” says Lane, “There's so much more progress that we can achieve with our team and I'm seeing it every day. It’s really something special.”

Find out more about ERM on the company’s website and LinkedIn.