AI Accountant

Abstract

This final year project suggests the conceptual design of an AI accountant software that makes use of cloud-based technologies, conversational artificial intelligence, natural language processing (NLP), and optical character recognition (OCR). Using straightforward chat phrases (such as "Paid rent Rs. 20,000") or by uploading images of receipts, the system is intended to enable SME owners, independent contractors, and microbusinesses to document financial transactions. Without requiring to hire a professional accountant, these unstructured inputs are automatically transformed into structured accounting entries, organised ledgers, and basic financial reports. This idea was basically derived by looking at the financial conditions of Pakistan and the microbusiness owner who were unable to afford a full time accountant and due to which they have to keep their records by themselves and later on they result in missed transactions, inaccurate profit and loss statement, no balance sheet and no vouchers for the tax authorities. The issue is further validated by primary research conducted through a survey of 22 SMEs, which reveals a widespread reliance on manual records, infrequent bookkeeping updates, heavy use of WhatsApp for business communication, and strong interest in a chat-based AI accounting assistant tempered primarily by worries about data accuracy and privacy. A conversational interface, an NLP engine, an OCR module, an automated ledger engine, a reporting and basic tax compliance module, and a security layer make up the modular architecture of the suggested system. High-impact features including conversational transaction entry, receipt scanning, automated bookkeeping, real-time financial reports, and streamlined GST aligned record organization are the main focus of the Minimum Viable Product (MVP). The solution minimises learning obstacles and promotes consistent recordkeeping by integrating smoothly with WhatsApp-based communication. The study shows that an AI-driven conversational accounting assistant is both technically possible and suitable for Pakistani SMEs, while being restricted to conceptual design and MVP-level specifications. According to the study's findings, this kind of solution can greatly strengthen financial discipline, lessen administrative work, increase tax preparedness, and aid in the formalisation of SMEs.