Mining News

How to ready your company for AI-driven ESG performance evaluation

The ESG tidal wave making its way through the global investment and business community in recent years is mainstreaming sustainability at an unprecedented rate. The rise of ESG (Environmental, Social, and Governance) global disclosure frameworks, rating agencies, as well as data accumulation and analysis service providers have all played a pivotal role in this rapid change.

No longer limited to mere high-level negative screening, ESG performance evaluation has evolved into a combination of quantitative and qualitative analysis impacting major investment decisions.

This is driven by the growing pool of investors who understand that ESG issues are material to business and who are seeking to understand how these issues link to corporate strategy, risk management, and operational success.

Upcoming event: Edumine Shift: Women In Mining – Exploring Points of Resistance

In fact, the inaugural ‘Responsible Capitalism’ survey in late 2021 found that 88% of institutional investors now see ESG factors as more important than traditional financial metrics when making long-term investment decisions. And that even if short-term returns look attractive, 79% believe a company with ESG failings should be avoided.

However, research has also shown that as little as 30% of investors find the ESG information currently provided by companies to be decision-useful for assessing materiality.

The landscape

Investors, analysts and companies alike continue to flounder, asking and responding to questions they hope are meaningful, while ranking and rating entities like Sustainalytics and ISS are overwhelmed with the sheer volume of data coming in. In its current state, insights provided by the ESG market remain unreliable for most investors. In part, this is because rating firms can’t seem to agree what ‘good’ performance looks like, with substantial differences in their proprietary methodologies leading to ESG scores that correlate poorly and that companies can ‘game’. But there are two other critical factors contributing to the unreliability of ESG insights, which the commercialization of artificial intelligence is promising to disrupt, and far sooner than you may think.

The first is the reliance on company disclosure. This subjective, self-reported content is not only limited and usually outdated, but also frequently inaccurate or incomplete, whether intentional or not, leaving analysts to work with patchy data that provides little insight into real and current ESG risks.

The second is that the bulk of analysis work is still being done by human analysts. This is incredibly resource-intensive and means that only a limited amount of data can realistically be analyzed. The use of individual human analysts within an organization attempting to comb through vast volumes of data also comes with the significant risk of human error, omission of key data, and personal bias.

The result is that, although the importance of ESG is growing exponentially, the current systems seeking to provide accurate decision-useful information are falling short. Given the growing influence of ESG data in investment decisions, countless data accumulation and analysis providers as well as tech startups are in a race to jump on this immense opportunity. As a wave of disruptive technologies change the face of our lives and our world as we know it, this field too will not remain untouched.

What’s changing

The collision of digital innovation with the world of ESG data is rapidly beginning to undercut traditional methods. Countless new software applications have emerged that are focused on strengthening internal data management systems to spit out clean report-ready data. This is streamlining and simplifying the process of pulling together the information that agencies and investors seek. However, this digital collision is also happening behind the scenes amongst those using ESG data to inform their rating and investment decisions about your company. Against the backdrop of global ESG standard and framework harmonization, the collision promises to make a major impact on the field.

While not yet deeply explored and studied in the ESG world, machine learning, data scraping, satellite analytics, natural language processing (big data, sentiment analysis) and other types of artificial intelligence (AI) technologies are increasingly being harnessed to improve ESG analysis and research. Simultaneously, a massive tech-driven transparency disruption is also underway, most obvious in the recent launch of the ambitious ESG Book platform, a free public good that enables anyone to both access and compare companies’ disclosed information.

More diverse and timely company news, announcements, investor presentations, published documents (such as policies) and filings are being dissected and integrated, blending financial and non-financial data sources to monitor changes in corporate ESG risk profiles. Parabole.AI, RSMetrics’ ESGSignals, Util, PACTA (Paris Agreement Capital Transition Assessment), RepRisk, Impact Cubed, SenseFolio, ClimateBert and more are already facilitating technology-enhanced company and portfolio assessment that measures ESG-related outcomes, impacts, and performance – and does so more accurately and in a more tailored manner than we have previously seen, providing data-seekers with increasingly concrete and objective information.

And it’s not just company sources. In our online world, there is an abundance of unstructured data that is an extraordinary source of signals of all sorts. Software programs are beginning to extract this so-called “alternative data” in multiple languages all over the world, including from websites, maps, satellite imagery, traditional news media, social media, NGO’s and industry watchdogs, trade and industry publications, and many other sources, to provide unprecedented 360 insights.

For example, air pollution, waterway or land clearing data from satellite images can be compared in real to with your GHG, water usage, or land disturbance/reclamation disclosures. Or sentiment analysis of geofenced social media commentary about your company can be compared to your social license claims.

This introduces new possibilities of rounding out the ‘inside-out’ view of company performance with an ‘outside-in’ view from countless external sources, for a more holistic and multi-stakeholder perspective than previously possible. As traditional ratings providers begin to move in this direction, some entities, like Truvalue Labs, are even experimenting with fully excluding company’s self-reporting to see how assessment outcomes compare.

What to expect

Expect more democratic and holistic data aggregation processes. Expect analyses to integrate higher volumes of data and broader data sets. And expect this to include a great deal of data not sourced from your company. This will make for a more well-rounded and robust analysis of ESG performance, but inevitably removes a certain element of control or influence from company hands.

As non-company-reported external data is increasingly incorporated into performance analysis, it will soon be commonplace to identify dubious or outright erroneous information in the blink of an eye, enabling almost instant appraisals of the accuracy or truthfulness of a company’s ESG claims. Such software can be applied to entire ESG funds and portfolios, as well as specific companies.

Resultant accusations of greenwashing or fraud could devastate a company’s credibility and reputation in our current climate of global distrust and rising performance expectations. It can lead to penalties and litigation. And it can negatively affect key stakeholder relationships, including not just rating agencies and investors, but also regulators, communities, and host governments that hold your current and future license to operate in their hands.

In this context, public sentiment from the local to the global, regarding any given company, stands to become more powerful than ever. And as society-wide expectations for both transparency and authenticity grow, PR spin will not suffice. Undertaking well-resourced, strategic efforts to develop (and deliver on) meaningful sustainability goals and embrace a role of genuine environmental stewardship will become key to company competitive advantage.

How to prepare

As fast as these ESG digital innovation trends have emerged, the rate at which they keep evolving shows no signs of abating. Public, regulator, employee, and investor expectations regarding strong sustainability performance and corporate transparency continue to grow. The bots will take some time to mainstream, but there is little doubt that their application will become global industry standard soon.

In the meantime, you can ensure your company is ready for it:

● Make sure your data is accurate and can be corroborated if compared to other external sources;

● Ensure your data can be found and understood by the people and the technologies seeking it;

● Remember that reporting and ratings are not ‘the work’ – focus on the quality and effectiveness of the operational ESG management approaches that produce the data you’re going to be judged on;

● Align your sustainability and ESG reporting with your official risk disclosures, as ESG-related risks are often poorly understood by many of the functions that could and should help manage them;

● Be strategic about where and how you disclose data on your corporate website(s), taking into account how deeply a data-scraping bot might have to dig to get to information;

● Ensure adequate and well-trained human resources are allocated to providing the information that investors and rating firms are asking for;

● Work directly with agencies where possible to make sure that the information they have is accurate and complete, and respectfully challenge any areas where you think data may have been misinterpreted or their assessment may be false, knowing that both human and machine error are always possible;

● Bring your technology and risk management teams into the ESG conversation to select appropriate software to accurately identify, understand, measure, manage and report on your material ESG risks;

● If deploying Mining 4.0 technologies to gather operational data, seize the opportunity to also begin collecting related and relevant ESG data too;

● If you don’t have a strong Environmental and Social Management System (ESMS) in place, develop one,including clear metrics and documented management practices and processes, which can be disclosed;

● If you haven’t undertaken a Materiality Assessment process, do so, and use it to inform a robust Board-endorsed ESG strategy.

Many serious questions and concerns remain about the digitalization of ESG data analysis processes, including around AI ethics, coded bias, cybersecurity, and data privacy. Their use will need to be well-monitored given such risks. However, the trend is not slowing down and may well entirely reshape the field of ESG performance evaluation.

As ESG tech continues to mature, smart leaders will move beyond the slog of annual sustainability reports to recognize and leverage the invaluable business intelligence that real-time well-rounded ESG data represents, to make ongoing business strategy and investment decisions. While laggard peers are likely to face unpleasant reputational damage and suffer serious credibility loss, prepared companies will be rewarded with the opportunity to differentiate themselves, winning with investors, regulators, insurers, analysts, partners, host communities and talent pools of the future.

Elizabeth Freele is a co-founder and managing partner of industry-focused sustainability think tank and ESG consultancy Sympact. She works with companies to help them ensure their social performance meets growing expectations.

Share this article