What is fine-tuning?
Fine-tuning is further training of an existing AI model on a smaller set of your own examples, so it handles a specific task or style more reliably.
Also calledfine tuning
Why it matters for a business
Fine-tuning costs time and data: you need many good examples, and you repeat the work when the base model changes.
More
It changes how the model behaves, not which current facts it knows.
For most company use cases, better prompts or RAG get results sooner. It pays off for narrow, high-volume tasks where a smaller tuned model can replace a larger, slower one.
What you need
A set of example inputs with ideal outputs, cleaned and checked by someone who knows the task.
Fine-tuning vs RAG
Use RAG when answers depend on documents that change; use fine-tuning for a fixed format, tone or classification.
In machine learning
The same idea applies to image and speech models: start from a trained model instead of training from zero.
- InputA question, a message, a document
- ContextYour data, found by search
- fine-tuningLanguage modelReads, reasons and writes
- ResultAn answer, or an action in your tools
What it costs
Starting prices of our services on this topic, in your currency.
Fine-tuning in practice
Further training of an existing AI model on your own examples so it handles one task in a set way.
Example
- A logistics company in Felixstowe sorts a high volume of customs documents into a few dozen types.
- A large model does it well but slowly.
- After fine-tuning a smaller model on labelled past documents, the sorting runs faster and costs less per document, and the team checks a sample every week.
Which term is it?
A large language model: an AI model trained on huge amounts of text to read and write language.

From the founder
I want AI that does a real job inside your tools, with a person checking what matters. We test it on your own data before you rely on it.
Anatoly MikhailovFounder of Pazl






