Overview

Engineered an autonomous FinTech pipeline powered by an AI Telegram Bot to ingest, parse, and process raw Ziraat Bank statements (XLSX/PDF) into a consolidated Google Sheets accounting core. The system fully automates multi-currency asset tracking, expense classification, and dynamic moving exchange rate calculations.

The Challenge

Manual transfer of hundreds of banking transactions from spreadsheets causes critical accounting errors. Key operational hazards include miscalculated volume-weighted asset purchase prices, forgotten transaction fees, and broken formula dependencies across historical balance sheets.

Tech Stack & Core Engineering

• AI Analysis & Deduplication Layer: Integrated Google Gemini AI via Telegram API to instantly clean raw inputs, remove duplicate records, and interpret transaction intents for deposits, withdrawals, and split-commissions. • Scenario Routing Core: Configured multi-branch logic in Make (Integromat) to execute real-time accounting rules, automate 99/1 cash desk fee allocations, and trigger cross-currency conversions (TRY to USDT). • Apps Script Automation Engine: Programmed a high-performance Google Apps Script backend to handle 1-click ledger history recalculation, automated formula injection (D2 ranges), and real-time moving average tracking without breaking legacy financial data.

The Result

• 100% automated parsing of unstructured multi-bank statements with zero manual entry overhead. • Instant history recalculation loop enabling seamless, 1-click past transaction corrections. • 24/7 transparent balance sheet visibility and flawless data integrity across integrated asset registries.