AI Support & Agents

Text Classification

定義

Text classification is the task of automatically assigning a category or label to a piece of text — such as tagging a support ticket by topic, language, or priority.

別名: text categorizationdocument classificationticket tagging

How it works

A model is given text and predicts one or more labels from a defined set. It can be trained on labeled examples (supervised learning) or, with modern LLMs, done zero-shot — you simply describe the categories in a prompt and the model classifies without task-specific training.

In support, classification powers automatic tagging, topic detection, language detection, and priority scoring. Intent recognition and sentiment analysis are both specialized forms of text classification.

Why it matters for support

Consistent classification is what makes ticket data usable: it feeds accurate routing, reliable reporting, and SLA prioritization. When AI tags tickets the same way every time, your dashboards and workflows can finally trust the labels — something manual tagging rarely delivers.

よくある質問

Is intent recognition a type of text classification?

Yes. Intent recognition classifies a message by the customer's goal, and sentiment analysis classifies it by tone. Both are specialized applications of text classification.

Can AI classify tickets without training data?

Modern LLMs can do zero-shot classification — you describe the categories in a prompt and the model labels the text without any task-specific training examples, though a few examples (few-shot) often improve consistency.

Text Classification

How it works

Why it matters for support

よくある質問

関連用語