Saudi Arabia’s digital-government agenda increasingly depends on decisions that must be fast, consistent, explainable and adaptable rather than merely digitised. This review examines how artificial intelligence, business rules and case management can be combined into an intelligent decision-automation architecture suitable for public services operating under Vision 2030. The paper synthesises peer-reviewed literature published from 2020 to 2025 across five intersecting streams: artificial intelligence in public administration, Saudi digital transformation, explainable and responsible artificial intelligence, hybrid decision management, and data-driven case management. The evidence indicates that artificial intelligence is strongest when used for prediction, classification, prioritisation and recommendation; business rules are strongest when policy, eligibility, thresholds and legal constraints require deterministic control; and case management is strongest when services involve incomplete information, exceptions, appeals and human judgement. Treating these capabilities as substitutes creates avoidable governance and operational risks. The review therefore proposes a layered framework in which trusted data and case context feed a decision-orchestration layer that coordinates artificial intelligence models, executable rules and case logic under human oversight. A seven-stage lifecycle—sense, validate, decide, explain, authorise, execute and learn—connects model outputs to accountable service actions and continuous improvement. The framework emphasises versioned policies, traceable evidence, appealability, bias monitoring, privacy controls and public-value metrics. The study contributes a practical architecture and governance model for Saudi public entities seeking to automate high-volume decisions without sacrificing legal compliance, institutional accountability or citizen trust.