Multi-Agent Collaborative Framework for Intelligent IT Operations: An AOI System with Context-Aware Compression and Dynamic Task Scheduling

The proliferation of cloud-native architectures, characterized by microservices and dynamic orchestration, has rendered modern IT infrastructures increasingly complex and volatile. This complexity generates overwhelming volumes of operational data, creating critical bottlenecks in information processing, task coordination, and contextual continuity during fault diagnosis and remediation. We propose AOI (AI-Oriented Operations), a multi-agent collaborative framework that integrates specialized agents with an LLM-based context compressor. AOI introduces a dynamic task scheduling strategy that adaptively prioritizes actions based on real-time system states, together with a three-layer memory architecture comprising Working, Episodic, and Semantic layers to optimize context retention and retrieval. Extensive experiments on synthetic and real-world benchmarks demonstrate that AOI achieves a 72.4% context compression ratio while preserving 92.8% of critical information, and significantly improves operational efficiency with a 94.2% task success rate and a 34.4% reduction in mean time to resolution (MTTR) compared to the best baseline. These results indicate that AOI provides a scalable and context-aware paradigm for autonomous IT operations.