Fine-tuning is further training an existing AI model on extra examples so it gets better at a specific style or task. Unlike a prompt or a saved skill, fine-tuning changes the model itself. It's more involved and usually a job for technical teams — and most businesses get what they need long before it's worth considering.
When people talk about "customising" an AI, fine-tuning is often what comes to mind — but it's actually the least common way most businesses end up personalising their AI use. Fine-tuning means taking an existing, already-trained model and training it further on a specific set of examples, so it becomes better suited to a particular style, tone or task.
The key distinction is that fine-tuning changes the underlying model permanently, whereas a prompt or a saved Claude skill guides the model's behaviour without altering it. That difference has real consequences: fine-tuning requires preparing a quality training dataset, running the training process, and testing the result — a meaningfully bigger undertaking than writing a good prompt.
For the overwhelming majority of small and medium businesses, that effort isn't necessary. A well-structured prompt, or a reusable skill built around your house style and past examples, delivers almost all of the practical benefit people are actually after when they ask about fine-tuning — with none of the technical overhead.
ExampleA large organisation with a very distinctive, consistent writing style across thousands of documents might fine-tune a model to reproduce that style automatically. A typical small business chasing the same outcome — "make the AI write like us" — usually gets there faster and more cheaply with a well-built skill or a saved prompt template using real past examples.
Start with a good prompt. If that's not consistent enough, build a reusable skill with examples of your best work. Only look at fine-tuning if you have a technical team and a genuinely large-scale, specialised need that those simpler options can't meet.
See also
Learn to use this on your own work
Our AI Training & Enablement program takes non-technical Australian teams from the terminology to real, working AI builds on the job they already do.
Book a call →