Resource-Efficient LLM Application for Structured Transformation of Unstructured Financial Contracts

TL;DR

This paper introduces a resource-efficient, template-driven framework extending CDMizer to convert complex legal contracts into standardized formats, demonstrating competitive performance with smaller open-source LLMs for financial contract automation.

Contribution

The paper presents an extension of the CDMizer framework that enables efficient, accurate transformation of legal contracts into CDM, using smaller open-source LLMs for cost-effective automation.

Findings

Achieves competitive accuracy with smaller LLMs

Ensures syntactic correctness and schema adherence

Offers scalable, resource-efficient legal contract transformation

Abstract

The transformation of unstructured legal contracts into standardized, machine-readable formats is essential for automating financial workflows. The Common Domain Model (CDM) provides a standardized framework for this purpose, but converting complex legal documents like Credit Support Annexes (CSAs) into CDM representations remains a significant challenge. In this paper, we present an extension of the CDMizer framework, a template-driven solution that ensures syntactic correctness and adherence to the CDM schema during contract-to-CDM conversion. We apply this extended framework to a real-world task, comparing its performance with a benchmark developed by the International Swaps and Derivatives Association (ISDA) for CSA clause extraction. Our results show that CDMizer, when integrated with a significantly smaller, open-source Large Language Model (LLM), achieves competitive performance in terms of accuracy and efficiency against larger, proprietary models. This work underscores the potential of resource-efficient solutions to automate legal contract transformation, offering a cost-effective and scalable approach that can meet the needs of financial institutions with constrained resources or strict data privacy requirements.