Agentic Assistant for 6G: Turn-based Conversations for AI-RAN Hierarchical Co-Management
Udhaya Srinivasan, Weisi Guo
TL;DR
This work introduces an agentic, turn-based assistant for AI-RAN hierarchical co-management, designating EngineerChat and UserChat atop a three-layer architecture (interface, intelligence, knowledge) and an in-house BubbleRAN ORAN emulator to enable AI service planning, deployment, and management. It presents the HATT-E evaluation framework to assess hierarchical, multi-turn agent performance across 50 ecologically valid scenarios, revealing an overall score of 0.6724 with strong performance on basic monitoring tasks but notable challenges in complex planning, synthesis, and factual grounding (hallucinations). The system demonstrates real-time capabilities (average ~13 s responses) and data-efficient tool use, suggesting potential OPEX reductions for small enterprises while highlighting the need for improved planning decomposition and grounding. Overall, the work provides a concrete path toward intelligent, human-in-the-loop AI-RAN orchestration, with a clear set of strengths and concrete areas for future improvement in reasoning, multi-modal integration, and end-to-end reliability.
Abstract
New generations of radio access networks (RAN), especially with native AI services are increasingly difficult for human engineers to manage in real-time. Enterprise networks are often managed locally, where expertise is scarce. Existing research has focused on creating Retrieval-Augmented Generation (RAG) LLMs that can help to plan and configure RAN and core aspects only. Co-management of RAN and edge AI is the gap, which creates hierarchical and dynamic problems that require turn-based human interactions. Here, we create an agentic network manager and turn-based conversation assistant that can understand human intent-based queries that match hierarchical problems in AI-RAN. The framework constructed consists of: (a) a user interface and evaluation dashboard, (b) an intelligence layer that interfaces with the AI-RAN, and (c) a knowledge layer for providing the basis for evaluations and recommendations. These form 3 layers of capability with the following validation performances (average response time 13s): (1) design and planning a service (78\% accuracy), (2) operating specific AI-RAN tools (89\% accuracy), and (3) tuning AI-RAN performance (67\%). These initial results indicate the universal challenges of hallucination but also fast response performance success that can really reduce OPEX costs for small scale enterprise users.