Curating Legislation through Artificial Intelligence

A use case for ML/AI for Corporate ESG & Sustainability

Customer: US Multinational Fintech PLC

Who is the Customer?

Provides software solutions for companies' tax needs, including tax determination, data and insights, compliance and reporting, & document management. The customer approached Spark to explore the use cases and applications of ML/AI in summarising ESG legislation and delivering those summaries in a feed format.

Impact Spark Made:

  • Created a user interface that maintains curated content and serves the search and results to the customer's users.
  • Integrated feature/function in the customer's planning application, including shared key architecture pieces and approaches.

Value Spark Delivered:

  • Proved the hypothesis and use case of ML/AI to curate ESG legislation.
  • Spark architecture ingested and processed content using ML, feeding raw summaries and producing human-readable content.
  • Tracer bullet (pre-MVP) that validated end-to-end Spark concepts across the customer's environment.

What the customer needed:

A use case for ML/AI in content curation.

The customer plans to expand its offerings into an adjacent RegTech vertical: ESG & Sustainability. In these verticals, various regulations must be captured, curated and summarised for use by the customer's applications and served to the customer's own users as a content clearing-house feature.

Spark Target Block Architecture.

Target Block Level Architecture-03

Want a copy of this Case Study?

Spark Target Tracer Architecture.

Tracer Bullet Block Level Architecture-02

Engagement Focus.

Curating legislative content opened the possibilities of leveraging AI to programmatically summarise content by identifying key points such as effective date, taxes, penalties and definitions and ensuring the information is searchable and linked to the original source content.

The engagement scope included creating a user interface to maintain the content and serve their customers' search results. The scope further provides for integrating this feature/function into the customer's planning application, including shared key architecture pieces and approaches.

Outcome & Results.

  • Critical to success is exploring the proof points for filtering, multi-source web scraping, and document aggregation with knowledge graphs, and ontologies.
    This R&D engagement allowed Spark to prove that it is possible to use ML & AI to produce effective summaries from legislative resources using Spark's proposed architecture that ingested laws and processes the legislation using ML feeding a raw summaries feed, which produced mock human-readable content after producing an RDF file.
  • A Tracer Bullet (pre-MVP) phase that validated end-to-end concepts in the customer's environment.

Our Work Makes You Look Good