What is Fine-Tuning?
Fine-tuning takes a trained model and continues training it on a focused dataset to shape its behavior: a support tone, a domain vocabulary, a task format. It is how companies customize models, and one of the ways assistant behavior changes between releases without a new base model.
Relevance to AI visibility
Consumer assistants are fine-tuned continuously for helpfulness, safety, and answer style. Those passes can shift how willingly a model names brands, how many options it lists, and how it hedges. When your mention rate moves on a model update, fine-tuning of the assistant is a likely cause, not anything you published.
What it is not
Fine-tuning is not something an outside brand can do to ChatGPT or Gemini, and no vendor can 'fine-tune your brand into' a public assistant. Influence from the outside runs through content, coverage, and retrieval.