A Framework for AI-Driven Legal Case Analysis and Grounded Decision Support Based on Integrated RAG and Agentic NLP
International Journal of Advanced Research in Science Communication and Technology · 22.11.2026 · 486
DOI: 10.48175/ijarsct-40955 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Advanced Research in Science Communication and Technology, online erschienen am 22.11.2026. Verfasst von Prathmesh Dawkar und 3 weiteren Autor:innen. Die Arbeit nennt 15 Literaturangaben.
Abstract (Originalsprache)
Massive amounts of textual and semi-structured data exist in legal domains in the form of court records, petitions, court judgments, depositions, and gazettes. However, the data suffer from incompleteness, heterogeneity, and fragmentation at different jurisdictions, with predominantly unstructured formats. Preprocessing such legal data for supporting judicious analysis and decision making (i.e., Legal Data Wrangling) is a tedious, time-consuming, and laborious task requiring expert knowledge and skills. Moreover, existing automated methods, including recent advances using general-purpose language models, are limited to performing specific tasks and thus unsuitable for judicious decision making in legal contexts. In this work, we propose a holistic idea and approach for an Automated Legal Case Analysis and decision support system called JuriumLegal++. This system employs an Agentic AI setup where an orchestrating AI agent, which is built upon a large language model (LLM) and a stateful graph (LangGraph), coordinates multiple specialized agents for profiling legal documents, cleaning legal statutes, standardizing legal terms and precedents through retrieval augmented generation (RAG), feature engineering for judicious decisions, validating evidence and generating explainable decision summaries. Document segmentation and similarity transformation were performed using deterministic database operations and vector calculations. If the quality of context alignment and accurate evidence is below a threshold, an in-built quality check forces the system to re-plan its workflow. The proposed methodology was designed and evaluated for an Indian Legal Document Corpus (ILDC) and High Court precedents and evaluated in 5 quality aspects and legal retrieval measures.
Zitieren
APA
Dawkar, P., Shejul, G., Shinde, S., & Patil, D. P. (2026). A Framework for AI-Driven Legal Case Analysis and Grounded Decision Support Based on Integrated RAG and Agentic NLP. International Journal of Advanced Research in Science Communication and Technology, 486. https://doi.org/10.48175/ijarsct-40955
Vancouver
Dawkar P, Shejul G, Shinde S, Patil DP. A Framework for AI-Driven Legal Case Analysis and Grounded Decision Support Based on Integrated RAG and Agentic NLP. International Journal of Advanced Research in Science Communication and Technology. 2026:486. doi:10.48175/ijarsct-40955
BibTeX
@article{Dawkar2026,
author = {Dawkar, Prathmesh and Shejul, Ganesh and Shinde, Sandesh and Patil, Dr. Pritesh},
title = {A Framework for AI-Driven Legal Case Analysis and Grounded Decision Support Based on Integrated RAG and Agentic NLP},
journal = {International Journal of Advanced Research in Science Communication and Technology},
year = {2026},
pages = {486},
doi = {10.48175/ijarsct-40955}
}
EUVIMED