Expense tracking is one of those critical workflows that businesses rely on—but almost everyone does poorly. Receipts get lost, entries are delayed, and finance teams operate without real-time visibility. To solve this, I built a Telegram-based expense tracking system that transforms everyday messages into structured financial records. Users can simply send a receipt photo or type a quick message, and the system uses an LLM to extract key expense details, validates the information through a human-in-the-loop confirmation step, and logs it instantly into a centralized ledger. Beyond simple logging, the system introduces real-time financial awareness. High-value expenses automatically trigger alerts to the CFO, while a built-in natural language query feature allows users to ask questions like “How much did we spend on transport this month?” and get immediate answers. The result is a seamless, intelligent workflow that replaces manual entry with fast, accurate, and auditable expense tracking—without requiring users to change their behavior.