Blog article
See all stories »

How AI and data enables ESG to make real world impact

ESG is a Data and AI problem

The benefits of incorporating Environmental, Social and Governance (ESG) within business targets are well understood by companies and regulators, and especially investors. Research shows a quarter of all fund investors planned to increase holdings in the sustainable sector over the next half year. 

Organisations looking to cement themselves as leaders in Corporate Social Responsibility (CSR) – in any sector, financial services included – must take technology-driven approach in order to enhance operational resilience and to satisfy investors. However, at the moment, ESG initiatives at most companies rely heavily on throwing bodies at the problem, resulting in manual and time-intensive processes that limit an organisation’s ability to respond to rapid changes in the economy, geopolitics or consumer behaviour. 

Today, most companies do not possess a centralized ability to collect, analyse and report ESG data. For ESG to properly take off and make the global impact it’s meant to make, data and AI need to play a central role in all aspects of ESG from collecting the data, surfacing insights to business leaders and serving as a guide to direct the company in a more sustainable direction. 

 

“Greenwashed” phenomenon 

Common practice today is for companies to issue comprehensive ESG disclosure only once a year. The primary reason for the infrequency is that it takes months for organizations to collect the required data for analysis disclosure. Simply locating the wide array of datasets, from carbon emission to diversity and inclusion (D&I) metrics, at a large company is a complex and time-consuming task on its own. As a result, most companies simply do not know how they are faring against their own stated ESG goals in a timely manner.

Given the diversity of the data involved, yearly disclosures (in the form of CSR reports) have become more marketing than disclosure. For example, a leading investment bank recently released it’s a 70-page annual sustainability report. Of those 70 pages, only 3 are dedicated to hard metrics. Do these disclosures items highlight the most important ESG performance metrics? Or are they simply the easiest observable data points? 

Without the ability to find, analyse and disclose ESG performance data in a timely manner, ESG efforts are turning into marketing exercises that sometimes lack solid grounding. This is well documented in what is called “greenwashing”, a term indicating institutions misleading consumers and/or investors about the performance of their Environmental, Social and Governance goals. 

 

Solving the Data problem first and foremost

Unlike financial data, ESG disclosure currently does not have generally accepted principles. In order to fix this, entrepreneurs and coalitions have stepped in to fill the gap. As of today, there are over 100 providers of ESG data that serve corporations and investors. Additionally, a patchwork of consortiums works to provide guidance on specific issues within ESG. 

While these developments are commendable, proper ESG standards will take years to consolidate. Given the current global situation today, companies and investors do not have the luxury to ‘wait and see’. Consumers are increasingly demanding sustainability from companies they purchase from, as are the majority of influential investors. 

In order to cement themselves as ESG leaders, companies and brands need to address foundational data problem: Volume, Velocity and Variety. 

  1. Variety – ESG data takes many different shapes and forms. A company needs to potentially collect hundreds of datasets from internal and external sources, that come in different formats and schemas.

  2. Velocity – ESG data constantly changes. Negative press articles, shifting consumer behaviour, product recalls, etc. The quicker you can analyse the data, the quicker a company can course-correct.

  3. Volume – Analysing data coming from social media, news, IOT devices at factories, etc quickly becomes a Big Data problem. A company might need to analyse terabytes of data on a daily basis to ensure there is no drift in their ESG performance or consumer perception of their ESG performance. 

If these foundational data problems are not solved, organisations cannot have an accurate understanding of their own ESG metrics, let alone deploy advance analytics/AI to draw insights and perform predictive analytics to enhance operational performance.

 

The corporate commitment to ESG

Once companies have a grasp on their own ESG performance, they must also focus on how they are being perceived by consumers, investors and the market. They need to reverse-engineer the ratings that an ESG vendors might assign to the company and have the ability to push back (with evidence) when needed.  They need to be data-driven in analysing the ESG performance of their supply chain, because a scandal at a single supplier might have significant consequences to financial performance and/or brand perception.  

For illustrative purposes, consider an Auto executive who wants to assess and verify the ESG performance of a manufacturer that makes its batteries. 

Today, he or she would have to rely on self-disclosure and/or third-party data vendor. This means that options are limited to either taking self-disclosure at face value or cross checking the disclosure with ratings from an ESG data vendor.

But tomorrow, ESG leaders can use AI and all types of data – structured, unstructured and alternative. To start, alternative data will be used to get insights on the environmental impact of a lithium mine which will enable the analysis of air pollution levels, water quality in nearby lakes and health records of nearby populations. Next, AI techniques such as embedding will help see how strongly environmental concepts are associated with the battery manufacturer in the media, shading light on any potential gap between what is said and what is done. Finally, Monte Carlo simulations can be run at scale to witness how various climate and economic conditions could impact the manufacturers ability to adhere to ESG standards. 

Data and AI are able to reinforce ESG compliance in ways that third party data vendors, auditors or rule-makers cannot. They help ESG become firmly integrated into business strategy and performance. They create a virtuous ESG Loop, where goal-setting, bench-marking and course-correcting reinforce sustainability. Ultimately, this what will empower ESG to make the real difference in the world it’s meant to make.

 

Author: Junta Nakai is the Global Industry Leader for Financial Services at Databricks. In his capacity, he is responsible for driving the world wide adoption of the Unified Data Analytics Platform across Capital Markets, Banking/Payments, Insurers and Data Providers. Prior to joining Databricks, Junta spent 14 years at Goldman Sachs.

a member-uploaded image
4648

Comments: (1)

Richard Peers
Richard Peers - ResponsibleRisk Ltd - London 05 August, 2020, 21:381 like 1 like

Great blog Junta, couldn't agree more

Junta Nakai

Junta Nakai

Global Industry Leader - Financial Services

Databricks

Member since

29 May

Location

New York City

Blog posts

1

This post is from a series of posts in the group:

Big Data

this Group Focusing Big Data Area and Data Engineering /Data Warehouse /Data Analytics


See all