Generative AI (GenAI) has become a transformative tool across many businesses including the networking sector. AI agents such as Retrieval-Augmented Generation (RAG) are being deployed to automate tasks, enhance customer interactions, and enable intuitive data exploration. Vendors are now embedding large language models (LLMs) into Network Management Systems (NMS), allowing users to manage infrastructure through natural language interactions. This has created new business opportunities for vendors, as many customers previously turned to open-source network management tools to avoid vendor lock-in —especially in wide area networks (WANs) managed by Internet Service Providers (ISPs).

‍

Notably, all major networking vendors have introduced AI-enhanced NMS platforms. Meanwhile, the open-source community is working to close the gap, with prototypes emerging from groups like the Network Automation Forum and National Research and Education Networks (NRENs). In the Netherlands, the Future Network Services (FNS) National Growth Fund project is exploring AI/Ops and digital twin technologies to simulate network changes before deployment. However, practical digital twin implementations remain challenging yet due to limitations in replicating hardware behavior and real-world traffic loads.

‍

AI-driven tools can boost operational efficiency, but organisations must weigh the risks of vendor lock-in versus investing in open-source skills and flexible architectures.