What role do algorithms play in assessing my company’s ESG performance and what can I do about it?

It’s not easy to evaluate a company’s overall environmental, social and governance (ESG) performance. Mandatory reporting is not required everywhere, self-reporting company data can be unreliable, and there are a wide variety of guidelines for disclosing ESG data. On top of that, there is wide disagreement on which factors are the most important in ESG. Some rating agencies look at value creation, while others look at risk. Some weigh carbon emissions more heavily, while other investors may be more interested in regulatory- or labor-related risk. Ultimately, the lack of a standardised rating and monitoring system for ESG performance leaves it a difficult to navigate.

For ESG investors, this ambiguity is both a problem and an opportunity. The lack of transparency and clearly-defined objectives makes it more difficult to determine which companies are most sustainable. This is a problem because high-performing ESG companies generally outperform their peers over the long term, making them exciting investment opportunities. On the other hand, the confusion provides an opportunity to gain an edge on the market if investors are able to better understand the signals that influence a company’s long-term value. With trillions of dollars invested in ESG-related funds, knowing who is a high ESG performer is a significant competitive advantage. Better ESG not only prevents negative impacts on the environment, but it creates an opportunity for companies to attract more investors, based on their core values.

This has resulted in a steep rise in the use of artificial intelligence (AI), machine learning (ML), natural language processing (NLP), and other advanced technologies to help evaluate ESG performance. These technologies are perfectly suited for these purposes because of the wealth of unstructured data, with varying degrees of relevance. It would be impossible for analysts to sift through thousands of reports, earning calls, news articles, social media feeds, third-party reports, etc. to determine a company’s overall sustainability performance, but a machine can analyze years of performance data in a matter of minutes.

Furthermore, the algorithms are less subjective and biased relative to their human analyst counterparts. Whereas individuals often make mistakes or come to very different conclusions based on the same information because of their biases, backgrounds and attention spans, algorithms are objective, data-driven and don’t get tired. AI is also able to identify hidden trends in company’s that are not always obvious. This helps investors and other external stakeholders when making investment decisions.

Therefore, investors and other external stakeholders use these AI tools for various purposes. Machine learning can help identify extreme stock price moves and the predictive mechanisms behind them. [Deep learning](https://www.ibm.com/cloud/learn/deep-learning#:~:text=Deep learning is a subset,from large amounts of data.) can identify patterns in a large, unstructured dataset and use them to build predictive models that help companies make better investment decisions. Natural language processing can look at earnings call transcripts, sustainability reports, news articles, social media, or any other “soft”, language-based data and determine the relevance or credibility of ESG performance factors, create sentiment analyses, or provide other human-like intelligence that would take years and significantly more resources to assemble for an analyst or consultant.

As a result, many of the world’s top asset managers use AI to evaluate ESG performance. There are also a wide variety of rating agencies, analysts, and 3rd-party providers that use algorithms to help inform their decisions. What investors and raters look at depends on their goals, but it’s clear that AI has become a deep and widely-used source of information for ESG professionals.

What investors look at

Asset managers sort through lots of data. AI helps investors identify patterns to understand where an industry is headed and how to find undervalued funds. It’s a huge part of the investor toolkit because it can help with operational efficiency and identify sources of value. For example, BlackRock uses AI to help analyze the thousands of quarterly earnings call transcripts and broker reports that they would otherwise need to have an analyst digest. But they also own part of ClarityAI, which they use to identify sustainable businesses.

Many investment firms use AI to assist with risk management. Since ESG data can tell you about how vulnerable a firm’s business model is to disruption, asset managers will often look at the sources of risk that are most material to the firm’s industry. Some ESG issues are clear-cut. For example, avoiding the oil and gas sector or heavy fossil fuel consumers is critical for a fund that is 20+ years from maturity. But what about a firm’s tenuous labor relations or presence of business operations in an unstable democracy? Many of the early signals of risk will come from sources such as news reports or even social media. Risk-averse investors such as pension fund managers will use data derived from AI to detect potential sources of trouble before they occur.

Company reports are also critical to investment decisions. What a company does (or does not) report on is very pertinent information. When a company omits or disguises key data points from its report, it is a clear cause for concern for investors. Deutsche Bank uses AI to weed out “greenwashing”, which is the practice of disguising unsustainable business practices with misleading information on how environmentally- and/or socially-friendly a company is. They also look at the indicators that are most material to the client’s industry and evaluate how much information is provided about the topic and how relevant the information is. Tools like AI can tell an investor how reliable and informative a report is.

Another key area where AI is used is for index funds or robo investors that use algorithms to determine the makeup of a fund’s portfolio. ESG data such as carbon emissions or overall ESG performance can be used as a screening tool or to identify high performing firms to invest in at higher levels.

Rating Agencies and Other 3rd Parties

Aside from investors, there is also an entire ecosystem of companies that attempt to determine companies’ ESG performances. Companies like Sustainalytics, MSCI, ISS, RobecoSAM, etc. attempt to give values to ESG performance by evaluating different indicators. While these firms all have different goals, scopes, methodologies and weightings for determining performance, they all use technology to drive their ratings.

For instance, 45% of MSCI’s data comes from AI-assisted datasets that do not rely entirely on corporate disclosures. Refinitiv uses AI to monitor news/social media and explicitly factors that into their overall ESG rankings. Their application specifically looks at outsider perceptions. RepRisk is more similar to a credit rating agency, and they use artificial intelligence to detect any sort of ESG-related risk. They use AI to detect the credibility and severity of ESG-risk incidents.

There are also data providers such as Act Analytics who provide investors with AI-assisted ESG news data that uses Natural Language Processing (NLP) to measure sentiment analysis of news articles. These can be adjusted to prioritize timeliness, volume, or credibility depending on the needs of the investor.

Regardless of the end user, all of these decisions are being made automatically, by algorithms. AI has huge potential to assist in optimising ESG performance through identifying and recommending actions for companies wanting to improve their performance. While humans will always be involved in investing, technology will be relied on more and more to determine the company ESG performance. Therefore, it has become critical for companies to ask themselves: How can we harness the power of AI to better improve our company’s ESG performance and perception? Companies that do not have a clear answer to this question risk wasting resources, being misunderstood and having investors choose their competitors over them.

How Eunoic Helps

A company’s value is affected both by its sustainable performance and being perceived as such. Eunoic provides resources to help in both ways. Using AI-infused applications, the company helps diagnose potential ESG issues - both in performance and perception - and provides clear and actionable ways to improve. With increasing expectations from investors, consumers, employees and other key stakeholders, sustainability is too important to get wrong. If you would like a free demonstration of our application, including sustainability report review, and news or social media sentiment analysis, please contact us for a free consultation.