Banking platforms operate in one of the most demanding environments for intelligent systems. Payments and core banking services must remain highly available, low-latency, secure, and auditable, even as organizations continue frequent software change, platform modernization, and tighter regulatory scrutiny. In this setting, traditional release governance and static monitoring approaches are no longer enough. They are often reactive, manually intensive, and poorly suited to the complexity of distributed banking systems.

This keynote presents a computational-intelligence approach to delivery resilience through Predictive Delivery Intelligence (PDI). Rather than treating software releases as routine pipeline events, PDI models each change as a risk-bearing decision point. It combines signals from code changes, CI/CD pipelines, runtime telemetry, and governance controls to produce an interpretable change-risk score and an evidence bundle that supports real-time release decisions. The approach brings together machine learning, explainability, policy-as-code, selective regression, and progressive delivery into a closed-loop decision framework designed for regulated environments.

The session will show how intelligent risk scoring can reduce noise, improve triage, and help engineering teams distinguish between normal operational variation and meaningful release risk. It also highlights how explainable models and audit-ready evidence can make AI-based decisioning trustworthy in financial services. Based on the paper’s replay-based evaluation, this framework demonstrated improvements in CI efficiency, rerun reduction, validation time, modeled recovery outcomes, and governance effort across a simulated 26-week environment spanning 240 services and 6,864 change controls. Attendees will leave with a practical blueprint for applying computational intelligence to software delivery resilience in modern banking systems.

https://theioes.org/conference/ijcaci2026/speaker/talk/106