Ben Howard-Cooper, Head of Sustainable Finance at Briink, explains how advancements in AI could help investors overcome common ESG reporting challenges.
The transition to a sustainable economy requires vast amounts of capital to flow into environmentally and socially positive activities at an unprecedented scale and pace.
Europe is leading the way with flagship regulations including the EU Taxonomy and CSRD, which provide a clear framework to define truly sustainable projects and mitigate “greenwashing”. However, the complexity of data disclosure requirements embedded in these regulations creates usability issues and limits the capacity for the market to comply. Primarily, there is an incredible amount of granular data that needs to be collected from a broad array of stakeholders, many of whom are simply not up to date on ESG frameworks. Thankfully, there are emerging technology solutions that can help manage these issues at scale, including Artificial Intelligence (AI) and in particular natural language processing (NLP).
NLP is the branch of AI concerned with the interpretation of text and speech data written or spoken in natural language, or human language. Data scientists, linguists and computer scientists working in the field of NLP build programs and algorithms that can understand human language and respond to it, summarize it, or re-elaborate existing text. Applications of NLP have seen a boom in recent years, including the much-hyped release of ChatGPT. Despite the many applications, at its core, NLP is best used to generate insights from unstructured data sources.
According to estimates, up to 90% of the data generated globally every day is unstructured. One type of data that comes very often in unstandardized form is sustainable finance data. The sustainability data needed to compile ESG KPIs for regulatory and stakeholder reporting lives across unstandardized reports and documents, which is a prime target for NLP solutions. Essentially, ChatGPT for ESG.
As a company that works at the intersection between AI and ESG, we at Briink are particularly familiar with this topic. Here we explore 3 ways NLP can accelerate sustainable finance:
Evidence extraction
Sustainable finance regulations, including the Taxonomy, include large sets of technical screening criteria to assess if an activity can be considered truly sustainable. NLP can help process millions of unstructured document sources, from ESG reports to life cycle assessments, on the hunt for evidence to support sustainability claims. The technology extracts the relevant passages from the documents and maps them to the appropriate legal criteria automatically, an otherwise highly manual and time-consuming process.
Differences in applied terminology across sectors and jurisdictions is a massive roadblock for standardized ESG reporting and data collection. The way a certain regulation refers to an economic activity can differ drastically from the way a given firm or investor classifies. NLP solutions can solve this issue by quickly sorting and assigning evidence based on broad meaning rather than specific keyword matching.
Summarization
The progress of NLP abstractive summarization over the last years (including OpenAI’s summarizer) has been groundbreaking. While there is still some way to go until this technology can deliver final results without human quality assurance, it holds significant promise for supporting sustainable finance usability with the prospect of rapidly ingesting and summarizing the deluge of complex legal documents. ESG regulations and frameworks touch an increasingly large part of the economy, with a greater share of companies and workers expected to manage ESG data collection and reporting processes for the first time, without much support. NLP solutions to simplify and contextualize regulation enables the market to comply, creating the necessary data pipeline to support greater transparency on sustainability metrics.
NLP solutions for summarization can also be deployed to track and simplify interconnected regulation. The EU Taxonomy references over 3,000 pages of directives and legislation that are essential to understanding the full scope of reporting requirements. Entity detection models can help link all of the referenced texts into one source of truth, making it easy for reporters to manage reporting.
Activity classification
The EU Taxonomy provides a detailed definition of over 100 sustainable activities. However knowing which, if any, apply to your company, portfolio, supplier or customer list can be a tricky and time consuming process. The regulation refers to businesses in ways that might not match internal lexicon. Further, companies can be engaged in numerous activities, or at times one activity can be classified under multiple Taxonomy listed activities. Using NLP models to scan large volumes of unstructured documents and find evidence of these activities automatically can significantly increase accuracy compared to a manual approach, while reducing the time by orders of magnitude.
In conclusion, while there has been ample attention on companies that are directly supporting climate change mitigation and adaptation efforts, there has been relatively little focus paid to the enabling activities. Firms that support the broad adoption of regulation, making it easier to bring innovative solutions to market, facilitating the flow of information to support data back decisions are an equally important part of the sustainability ecosystem. To get to the level of data disclosure and standardization required to power effective capital allocations we will need to leverage technology driven solutions.
On the back of these opportunities, the market for ESG regulatory reporting solutions has recently taken off. We expect partnerships between like-minded firms will accelerate the use of AI assisted technologies across the sustainability sector.
All in all, the increasing regulatory and stakeholder scrutiny on ESG factors has done exactly what it was intended to do, which is move sustainability away from the margins of the conversation and require that markets put non-financial impacts at the center of decision making. Of course, such large scale change delivered so rapidly creates the space for confusion and limits the effectiveness of the frameworks. Deploying emerging technologies to manage ESG data can enable it to get there, and is the starting point for the development of more groundbreaking AI applications
Ben Howard-Cooper, Head of Sustainable Finance at Briink, explains how advancements in AI could help investors overcome common ESG reporting challenges.
The transition to a sustainable economy requires vast amounts of capital to flow into environmentally and socially positive activities at an unprecedented scale and pace.
Europe is leading the way with flagship regulations including the EU Taxonomy and CSRD, which provide a clear framework to define truly sustainable projects and mitigate “greenwashing”. However, the complexity of data disclosure requirements embedded in these regulations creates usability issues and limits the capacity for the market to comply. Primarily, there is an incredible amount of granular data that needs to be collected from a broad array of stakeholders, many of whom are simply not up to date on ESG frameworks. Thankfully, there are emerging technology solutions that can help manage these issues at scale, including Artificial Intelligence (AI) and in particular natural language processing (NLP).
NLP is the branch of AI concerned with the interpretation of text and speech data written or spoken in natural language, or human language. Data scientists, linguists and computer scientists working in the field of NLP build programs and algorithms that can understand human language and respond to it, summarize it, or re-elaborate existing text. Applications of NLP have seen a boom in recent years, including the much-hyped release of ChatGPT. Despite the many applications, at its core, NLP is best used to generate insights from unstructured data sources.
According to estimates, up to 90% of the data generated globally every day is unstructured. One type of data that comes very often in unstandardized form is sustainable finance data. The sustainability data needed to compile ESG KPIs for regulatory and stakeholder reporting lives across unstandardized reports and documents, which is a prime target for NLP solutions. Essentially, ChatGPT for ESG.
As a company that works at the intersection between AI and ESG, we at Briink are particularly familiar with this topic. Here we explore 3 ways NLP can accelerate sustainable finance:
Evidence extraction
Sustainable finance regulations, including the Taxonomy, include large sets of technical screening criteria to assess if an activity can be considered truly sustainable. NLP can help process millions of unstructured document sources, from ESG reports to life cycle assessments, on the hunt for evidence to support sustainability claims. The technology extracts the relevant passages from the documents and maps them to the appropriate legal criteria automatically, an otherwise highly manual and time-consuming process.
Differences in applied terminology across sectors and jurisdictions is a massive roadblock for standardized ESG reporting and data collection. The way a certain regulation refers to an economic activity can differ drastically from the way a given firm or investor classifies. NLP solutions can solve this issue by quickly sorting and assigning evidence based on broad meaning rather than specific keyword matching.
Summarization
The progress of NLP abstractive summarization over the last years (including OpenAI’s summarizer) has been groundbreaking. While there is still some way to go until this technology can deliver final results without human quality assurance, it holds significant promise for supporting sustainable finance usability with the prospect of rapidly ingesting and summarizing the deluge of complex legal documents. ESG regulations and frameworks touch an increasingly large part of the economy, with a greater share of companies and workers expected to manage ESG data collection and reporting processes for the first time, without much support. NLP solutions to simplify and contextualize regulation enables the market to comply, creating the necessary data pipeline to support greater transparency on sustainability metrics.
NLP solutions for summarization can also be deployed to track and simplify interconnected regulation. The EU Taxonomy references over 3,000 pages of directives and legislation that are essential to understanding the full scope of reporting requirements. Entity detection models can help link all of the referenced texts into one source of truth, making it easy for reporters to manage reporting.
Activity classification
The EU Taxonomy provides a detailed definition of over 100 sustainable activities. However knowing which, if any, apply to your company, portfolio, supplier or customer list can be a tricky and time consuming process. The regulation refers to businesses in ways that might not match internal lexicon. Further, companies can be engaged in numerous activities, or at times one activity can be classified under multiple Taxonomy listed activities. Using NLP models to scan large volumes of unstructured documents and find evidence of these activities automatically can significantly increase accuracy compared to a manual approach, while reducing the time by orders of magnitude.
In conclusion, while there has been ample attention on companies that are directly supporting climate change mitigation and adaptation efforts, there has been relatively little focus paid to the enabling activities. Firms that support the broad adoption of regulation, making it easier to bring innovative solutions to market, facilitating the flow of information to support data back decisions are an equally important part of the sustainability ecosystem. To get to the level of data disclosure and standardization required to power effective capital allocations we will need to leverage technology driven solutions.
On the back of these opportunities, the market for ESG regulatory reporting solutions has recently taken off. We expect partnerships between like-minded firms will accelerate the use of AI assisted technologies across the sustainability sector.
All in all, the increasing regulatory and stakeholder scrutiny on ESG factors has done exactly what it was intended to do, which is move sustainability away from the margins of the conversation and require that markets put non-financial impacts at the center of decision making. Of course, such large scale change delivered so rapidly creates the space for confusion and limits the effectiveness of the frameworks. Deploying emerging technologies to manage ESG data can enable it to get there, and is the starting point for the development of more groundbreaking AI applications
Share via:
Recommended for you