There are two main ways to make a general model fit a specific need. Fine-tuning further trains it on examples so the behaviour is baked in, while prompting simply gives careful instructions and context each time. Each suits different situations.
Prompting, often with relevant documents supplied at the moment, is quick, cheap and easy to change, and it covers many needs. Fine-tuning helps when a consistent style or specialised skill is required and enough good examples exist, but it costs more and must be redone as needs change. Teams increasingly start with prompting and fine-tune only when it clearly pays off.