83% Faster, 10x Fewer Errors: Engineering a Self-Healing Payment Saga Tracker

Abstract

This paper describes the re-architecture of a distributed payment system around an event-driven saga pattern, tracking transactions from initiation through settlement and automatically triggering compensations when a step fails. The results are given directly: transaction processing time fell by 83 per cent, payment errors dropped tenfold, and security incidents fell by 95 per cent. The pattern's value is that it converts fragile, error-prone distributed transactions into systems that recover without intervention, which matters most in payments, where a partial failure left uncompensated is a financial discrepancy rather than a technical one. The paper also covers a companion project, an Azure OpenAI-powered chatbot using retrieval-augmented generation and vector embeddings, which now handles between 50 and 70 per cent of customer support queries and reduced support costs by 10 to 15 per cent. Taken together, the two show how reliability engineering and AI-powered support compound, improving operational efficiency and cost at the same time.