---
title: "Why Fine-Tune?"
description: "A support router writes paragraphs instead of labels. Diagnose the failure first, weigh prompting, retrieval and fine-tuning, then run a small experiment."
canonical_url: "https://fanout.sh/ai/lessons/fine-tuning-why-finetune"
md_url: "https://fanout.sh/ai/lessons/fine-tuning-why-finetune.md"
access: "pro-preview"
---

# Why Fine-Tune?

A support router writes paragraphs instead of labels. Diagnose the failure first, weigh prompting, retrieval and fine-tuning, then run a small experiment.

- Access: Fanout Pro (public preview)
- Module: How to Fine-Tune Models

## A glimpse of Why Fine-Tune?

Suppose a support team wants a model to route messages to billing, delivery or account. A customer writes “I paid twice,” and the model replies with a helpful paragraph. The explanation may be sensible, but the routing system needs a label it can use.

Before changing the model, make the task explicit: “Return one of these three labels.” Add a few examples, including “I paid twice” → billing, and test messages the model has not seen in that prompt.

## Continue learning

Pro unlocks guided learning, complete notes, practical walkthroughs, exercises, and every AI Research lesson.

---
This representation contains public Fanout content only. Protected Pro lessons, account data, billing, checkout, and pricing are not included.

Browse the public content map: https://fanout.sh/sitemap.md