Intent categories
Intent categories are the set of reasons customers reach out to you — for example “billing question”, “cancellation request”, or “bug report”. When a message arrives, the AI assigns it to one of these categories. Well-chosen intents make your routing and your reporting far more useful, because every ticket is sorted by why the customer wrote in, not just who they are.
You manage them under Settings → AI → Custom intents. Each intent has a name, a slug (generated from the name), a short description of what it covers, and a few example messages — real phrasings a customer might use. Those examples are what teach the AI to recognize the intent, so the more representative they are, the more accurate classification becomes. Once defined, an intent becomes available as a condition in rules so you can route and act on it automatically.
Why intents matter
- Routing — send each intent to the team or group best equipped to handle it, so tickets land in the right place automatically.
- Reporting — see which reasons drive the most volume, spot trends over time, and staff accordingly.
- Consistency — every agent and every automation works from the same shared vocabulary.
Designing good categories
A strong set of intents is small enough to stay meaningful but complete enough to cover the bulk of your traffic. Keep these principles in mind:
- Use language your team already speaks — name intents the way you’d describe them in a stand-up.
- Keep categories distinct, so a single message rarely fits two of them equally.
- Cover your common requests first; you can always add niche categories later.
- Review your intents periodically and retire ones that no longer earn their place.
Keeping intents current
Your product and your customers change, and your intent taxonomy should follow. When you launch a new feature or notice a recurring request that doesn’t fit any existing category, add one. As you refine the list, the AI’s classifications — and the routing and reports built on top of them — get sharper.