Fine-tuning

Definition: Additional training applied to an already-built model, using your own examples, so it consistently adapts to a specific task, tone, or format — different from prompting, which shapes a single response rather than the model's underlying behavior.

Why it matters: Prompting has to be repeated and adjusted every time; fine-tuning bakes the adaptation into the model itself, so you stop needing an elaborate prompt to get the same consistent style or format every time.

Example: Fine-tuning a model on a year of past support tickets so it automatically answers in your team's actual tone, without needing that tone described in every prompt.