ARCHIVES

Recent advances in Large Language Models (LLMs) have created new opportunities for automating network management tasks. This paper presents an LLM-based intelligent assistant that translates natural language user requests into vendor-specific network configuration commands. The proposed system integrates prompt engineering, an LLM inference engine, validation mechanisms, and network automation frameworks to generate accurate configurations for platforms such as Cisco and Juniper. Implemented using Python with network automation tools, the assistant supports automated configuration generation and iterative refinement based on user feedback. Experimental evaluation demonstrates that advanced LLMs, particularly GPT-4, produce accurate and context-aware configurations while reducing manual effort and configuration errors. The study also discusses practical challenges, including prompt quality, model hallucinations, security considerations, and API costs. The proposed approach highlights the potential of generative AI to simplify network administration and provides a foundation for future research on context-aware, multi-vendor intelligent network automation.