There is no one-size-fits-all when it comes to Large Language Models (LLMs) in the context of service provision and O&M (Operations and Management). The same can be said of the overall LLM strategy. As operators deploy 5G Standalone (SA) with more complex cores, a key question becoming more urgent is whether to build fully customized LLMs incorporating their own data and exploiting domain expertise? The alternative is to harness massive investment made by others such as OpenAI with ChatGPT, Meta with Llama, or Google with its Gemini family. In the latter case, operators can still tune the models and bring in their own data, so that approach at first sight would surely appeal to many operators, especially smaller ones lacking…