Skip to content

🧩 Project Overview ​

Title ​

AI-Powered Taxation Assistant: Financial-Statement & Competitor-Based Tax Insights (RAG + LLM)


Motivation ​

Tax laws are intricate, and compliance often suffers due to:

  • Complex tax classification of financial items
  • Evolving Finance Acts and CBDT circulars
  • Lack of automated benchmarking with peer companies

This project aims to bridge the gap between raw financial data and actionable tax intelligence through AI-powered automation.


Objectives ​

  1. Build a Retrieval-Augmented Generation (RAG) pipeline that integrates Finance Acts, CBDT notifications, and company filings.
  2. Classify assessee tax regimes (individual/company; presumptive or regular).
  3. Generate personalized tax compliance suggestions.
  4. Compute a risk-of-notice score based on anomalies or discrepancies.
  5. Validate outputs through expert evaluation and user studies.

Expected Deliverables ​

  • Functional RAG+LLM prototype
  • Finance Act summary report for recent FYs
  • Evaluation metrics for accuracy and usefulness
  • Recommendations for scaling and future integration with tax APIs