FIS Worldpay has deployed an AI-powered authorization optimization system that dynamically routes transactions across multiple acquiring banks and payment networks to maximize approval rates. In its first 90 days of production, the system recovered $1.4 billion in transaction volume that would have been declined under static routing.
The optimizer uses reinforcement learning to determine — in real time and per-transaction — which acquiring bank, card network, and submission format combination has the highest probability of approval. It considers issuer-specific patterns, time-of-day effects, BIN-level behaviors, and historical decline reasons.