> ## Documentation Index
> Fetch the complete documentation index at: https://documentation.lobbychat.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# How to Test Your Hotel Chatbot Prompt

> Evaluate accuracy, tone and boundaries with realistic questions

Use **Test your chatbot** in your Lobbychat dashboard to check how your instructions work before and after changes.

## Test different guest situations

| Scenario | Example question | What to check |
| - | - | - |
| Routine information | “What time is check-in?” | Accuracy and concision |
| Specific needs | “Do your rooms have connecting doors?” | Evidence and uncertainty |
| Follow-up conversation | “And is breakfast included?” | Context and relevance |
| Staff action | “Can you change my reservation?” | No unsupported promises |
| Unknown details | “Can you guarantee an early check-in?” | Appropriate referral |
| Language variation | Ask a common question in another language | Clarity and consistency |

## Five checks for every response

1. **Accuracy:** Does it match verified hotel information?
2. **Relevance:** Does it answer the question asked?
3. **Tone:** Does it fit your hotel's service style?
4. **Clarity:** Is it understandable without unnecessary detail?
5. **Boundaries:** Does it avoid invented facts and uncompleted actions?

## After a change

Save the prompt, select **Retrain Chatbot**, wait for confirmation, then repeat the question. Saving alone does not complete retraining. Test related questions as well: a fix to one response can affect others.

See [How to Test Your Hotel AI Chatbot](https://documentation.lobbychat.ai/manage/test-chatbot.md) for the dashboard walkthrough.

**Next:** [Improve your prompt using chat logs](https://documentation.lobbychat.ai/prompting/improve-from-chat-logs.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.