For the complete documentation index, see llms.txt. This page is also available as Markdown.

Type or dictate a prompt

Enter a request as text or use the microphone in the assistant's message composer.

AI Assistant message composer with the microphone and send arrow

The microphone is beside the send arrow at the lower right of the composer.

Type a request

  1. Open the relevant chat.

  2. Click the message input and describe one task.

  3. Include the page, data source, and expected result.

  4. Review your wording, then use the send arrow.

Dictate a request

  1. Click the microphone beside the message input.

  2. Allow microphone access in your browser if prompted.

  3. Speak your request clearly, including the page and field names.

  4. End dictation using the recording control.

  5. Review the resulting text, correct any transcription errors, and send the request.

For example:

On the Tickets page, show Priority next to Status. Keep the existing sort order.

Check the result

Confirm the request names the correct fields before sending it. After generation, inspect the changed page.

If the microphone does not work, check browser microphone permission and the selected audio input. You can always type the same request. Spell out unusual table or field names in the text input if dictation mishears them.

Next: Manage chats and context.

Was this helpful?