You can ask Claude to classify inputs as follows.

Claude is a conversational agent, so often it will explain at length why it selected a given option. If you just want a fast multiple-choice answer and no explanation, you can put words in Claude's mouth and have Claude respond in a particular way.

For example, if the available options start with `(`, you can end your prompt with an Assistant message "`My answer is (`". That way you can immediately get a response in the first token.

UserYou are a customer service agent that is classifying emails by type. Email: \<email> Hi -- My Mixmaster4000 is producing a strange noise when I operate it. It also smells a bit smoky and plasticky, like burning electronics. I need a replacement. \</email> Categories are: (A) Pre-sale question (B) Broken or defective item (C) Billing question (D) Other (please explain)
AssistantMy answer is (
AssistantB) Broken or defective item

**Note:** You can also rewrite this prompt using a system prompt. It would involve moving everything in the `User` role except the email into the `system` parameter. See [system prompt prompting techniques](🔗) for more details.