A novel cloud-enabled enterprise information systems framework for accelerating digital transformation strategies

The co-evolution of micro-frontend UIs and backend microservices in cloud-native environments is complicated by UI–API semantic drift, dynamic workloads, and safe-deployment constraints. A tripartite agentic AI framework is introduced that combines reinforcement-learning (RL) agents for adaptive runtime control, large-language-model (LLM) synthesizers for automated artifact generation, and policy-as-code governors for security and compliance enforcement. The co-evolution process is formulated as a drift-aware Markov decision process, while Kubernetes-native traffic shifting supports progressive rollout and rollback. KubeSim was used for offline policy training; the reported performance results were obtained from live AWS EKS, Google GKE, and Azure AKS clusters. Across five benchmarks, the framework achieved a mean resolution success rate of 91.4%, compared with 62.0% for rule-based control, 71.1% for a controlled zero-shot LLM baseline, and 79.0% for partial RL. The framework also reduced normalized mean time to recovery by 80%, normalized change failure rate by 75%, and CPU utilization by 30 percentage points, while attaining a 98% audit pass rate. Component-removal and counterfactual-return analyses indicate complementary contributions from adaptive control, artifact synthesis, and policy enforcement. These results support the effectiveness, scalability, and governability of the framework for cloud-native enterprise digital transformation.