ESG AI – Harness the power of ESG data

GreenFI IMG 004

Every business spends money on ESG. Sustainability spend include consulting & advisory, software & data subscriptions, employee enablement, offsetting projects and a number of other areas in which a typical business spends money. Across businesses, managing ESG decisions is incredibly painful because of the siloed and fragmented nature of the data and software used.

Historically, tools in this domain have fallen into several buckets:

  • Standalone software workflows that don’t touch transactions
  • Reporting systems that address one need without any support for associated workflows in the business
  • Siloed workflow-focused systems that integrate with the related data collection rails

The ecosystem of tools to solve ESG management evolved this way primarily because they are either front ends to specific data gathering rails used by businesses, or for a specific business process. No primary stakeholders in the decision management process would design a siloed and fragmented system like this. They just don’t think about the lifecycle of making decisions this way.

In practice, the stakeholders care about three simple steps:

  1. Make sure there is a good process to control data sourcing & data maintenance.
  2. Aggregated information flow with relevant decision insights traceable to data, assurance and frameworks.
  3. Correctly categorize by user journey, provide decisioning insights and record it in a unified dashboard, ideally in a completely automated way.

But, what they have to deal with in the traditional siloed system is:

  1. An archaic and periodic data re-engineering process which is an enormous amount of low-value, manual work that involves accessing data from each system, categorizing it, recording it in the database, chasing people throughout the company for additional information about how to categorize specific information, dealing with a deluge of ESG reports and reconciling the flow of data through multiple systems. Simply put, a large part of the work undertaken by stakeholder teams reflects the limitations of tools they are forced to use.
  2. A lack of high-quality control and visibility into how ESG decisions are made by stake-holders across the business because they don’t have access to all the systems, or they don’t know how to interpret the data and they have to wait for 2-3 weeks after for the information to be collated as part of a transaction decisioning process. Most businesses are not enabled to make good decisions based on real-time ESG-related data.

Why hasn’t anybody solved this problem so far?

The primary reason the landscape of ESG tools has remained in silos until recently is due to the lack of maturity of the underlying ESG ecosystem and regulations. It’s important to understand that solving the ESG data and decision management problem the right way involves deeply integrating workflow-driven software and decisioning insights into an integrated platform.

Over the last 2 years, two big changes have made it easier (not easy!) to build systems that integrates the ESG workflows with AI:

  1. The maturity of ESG regulations, frameworks and standards are evolving at this point.
  2. Banks & Corporates have gotten better at allowing start-ups and earlier stage companies to integrate into their systems to provide the necessary building blocks to manage ESG and make sustainable investments.

This has allowed anyone who deeply understands the lifecycle of ESG investments in a business – starting with collaboration between team members as they decide how to invest or select suppliers, and ESG workflows for how it’s all categorized and recorded correctly in a ledger to bring the software workflows and investments together in a way that very closely maps to the actual workflows that happen inside a business.

Solving the global ESG data management problem is a huge opportunity

The trend of making globally distributed investments and trade relations has been accelerating over the last few years, but the pandemic brought it forward by 10 years. Companies at every stage of growth are now doing business around the world, which means the ESG management problem is global. If dealing with 5-6 systems to manage ESG is hard, adding 2-3 additional systems per international entity and meeting the needs of a distributed counter-party ecosystem is a big challenge for stakeholders. It’s a huge opportunity to deliver a single global platform to manage ESG data and decision insights for a business across locations.

Workflow-driven ESG AI

ESG data is an important enabler, but just that – an enabler. Solving a data engineering problem is the first step, but also workflow, process, information exchange and decisioning problems in businesses are critically important. Data is a utility, and the value is in the deep software layer above it primarily in the AI enabled ESG workflows.

Image 2 1024x699

ESG AI is a transformative technology

ESG AI technologies are based on years of research made in many parts of the industry and academia. The final breakthroughs, i.e. scaling training on aggregated data and aligning models with human feedback, finally made these technologies usable by many; these breakthroughs are made by very few actors. This kind of product has been in the making by only a few small teams across the world — the few ESG researchers are now the limiting factor to create new economic and sustainability actors in the field.

Most of the value in the emerging ESG AI market will be located in the hard-to-make decisioning work-flows, i.e. the generative models themselves. Those models need to be trained on thousands of sustainability scenarios, on millions of data points coming from high quality sources, which is one factor that sets a high barrier to entry. The second important barrier lies in the difficulty to assemble an experienced team, something that we are in a unique position of doing.

Current ESG providers do not meet market constraints

Many of the current ESG data providers have embraced a closed and narrow technology approach, which will dramatically reduce their market reach. In that approach, the model is kept secret and is only served through a rating. This raises the following important concerns for businesses:

Only exposing the output of models, instead of exposing the model entirely, makes it harder to connect with other components (data sources, retrieval databases, structure inputs, incidents combinations, integrated workflows, risk models, policy mapping and composed adjustments). The data used to train the model is kept secret, implying that we rely on a machine that has unidentified sources, and can produce uncontrollable outputs. Filtering efforts to address this issue are only a slim and breakable guarantee that the model will not output trustable content on which it may have been trained. As of now, this issue formed the basis of scrutiny and regulatory enforcement of many ESG data providers.

Businesses wishing to embrace ESG data are forced to feed their valuable business data and sensitive user data to a black-box model, typically deployed in the public cloud. This creates trust issues: models kept secret cannot be inspected to guarantee their outputs to be accurate, thereby preventing them to be deployed in business-critical decisioning applications. It also raises legal problems, in particular the one of falling under extraterritorial reach when sending data out of a company’s legal territory.

ESG ratings, ESG reports & Carbon accounting are commodity products

With the availability of mature Carbon measuring, ESG reporting and rating platforms, there is a glut of ESG software offerings in the market and this will only continue. Trying to acquire customers on the promise of reports, ratings, rewards and size of the line of pricing is a race to the bottom. But, that’s not the best way to maximize long term value creation. ESG management of the future are primarily AI enabled workflow-driven software products where the real value is in deeply understanding the entire lifecycle of how businesses make decisions based on data and to support that in software workflows. Also, ESG reporting is just one piece of the larger problem businesses have to solve.

ESG workflow model development

Image 3 1024x774

The use of AI technologies to recommend ESG decisioning can test a customer-transaction-portfolio-supply chain at model-scale, so it allows to test lots of data enabled scenarios repeatedly. You can refine the models over lots of workflows to incrementally improve ESG traceability and visibility. In recent years, regulators have imposed ever-more stringent regulations to be ESG compliant. All of this means that it is crucial for businesses to be efficient as possible and they get data as quickly as possible, as that enables them to get the correct decision recommendations.

The community of sustainability experts, data scientists, modelers, and software developers are building the Data Commons, a federated library of libraries of ESG factor data, plus analytics tools to derive the actionable decision metrics crucial for asset allocation, portfolio construction, incident analysis, risk analysis, supplier engagement, strategic planning and transition investment by corporates, and financial sector supervision. So effectively we are going from the place of ‘here is an alert for you to think about, what are you going to do about it to here is a solution or set of recommendations or staged scenario when this alert happens, we will evaluate it if you approve or reject it. We will work with staged set of edits and can apply blended intelligence meaning have a human evaluate it, thresholding and learning from it and let LLM assist / decide on recommended actions. Ultimately ESG AI can be designed to help investment managers, sustainability managers, procurement officers, facility managers, and other stake holders fine-tune decisions to drive reductions.