AI-Powered Tax Calculation and ITR Suggestion System

Abstract

Filing Income Tax Returns (ITR) in India can be confusing and time-consuming due to frequent changes in tax rules and the need to manually interpret financial data. Most accounting tools can store records but do not help users understand which tax rules apply, which ITR form to choose, or how their final tax amount is calculated. This paper presents an AI-Powered Tax Calculation and ITR Suggestion System that combines Retrieval-Augmented Generation (RAG) and Large Language Models (LLMs) to make tax filing simpler and more transparent. The system reads financial data directly from Tally XML files, analyzes income and expenses, and applies the latest tax regulations using a hybrid approach that blends rule-based logic with intelligent knowledge retrieval. It also generates clear, human-readable explanations describing tax liability, deductions, and the recommended ITR form. Testing with datasets showed accurate results for individuals, firms, and companies. The proposed system reduces manual effort, ensures up-to-date compliance, and improves user confidence through easy-to-understand AI-generated reports.