Researchers propose Agentic Nesting, a multi-agent framework to integrate legacy enterprise systems

Researchers from multiple institutions propose Agentic Nesting, a multi-agent collaboration framework designed to integrate heterogeneous enterprise applications by treating each system as an autonomous AI agent within a layered hierarchy.

The approach contrasts with conventional integration methods—such as Enterprise Service Bus (ESB), API gateways, and Robotic Process Automation (RPA)—which the authors argue suffer from high architectural coupling, escalating operational and maintenance costs, and limited intelligence capabilities.

In the proposed framework, each legacy application is encapsulated as a digital agent proxy that enables natural-language interaction and autonomous manipulation, while a central orchestrator coordinates task decomposition and dynamic dispatching across agents.

The methodology introduces two core concepts: the 'Application-as-Agent' integration paradigm and the 'Conversation-as-Integration' interaction philosophy, aiming to unify cross-application querying and process orchestration through a conversational interface.

The authors suggest the framework’s generalization potential extends to scenarios involving heterogeneous system coordination and large-scale data applications, though they do not provide empirical benchmarks or case studies in this paper.