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

# Ask your agent a question

<Info>
  Make sure you've navigated to a specific Chat (RAG) agent via the `Agents` list from left nav, or by typing the agent name in the `cmd+k` shortcut
</Info>

## Run your query

1. Type a question into the chat window and tap `enter` or the 'send' button to send it to your agent.

   <Tip>
     Your query can use a specific source to use for its response by passing in the sources `nickname` (see [Add Sources](https://docs.usesolari.ai/essentials/add_sources) to understand `nicknames`). \
     \
     If no source is mentioned, Solari's `model router` will choose the right model for the question based on the source description, contents, and previous questions.
   </Tip>

<AccordionGroup>
  <Accordion title="No source mentioned">
    1) If no source is mentioned, Solari's `model-router` guesses which source is most relevant based on the source contents, description and previous questions. 
    2) Solari wll return with the source it has chosen. The agent will use this source in 5 seconds, unless the user manually changes the source
       1. To change the source tap the message to open the source menu. Select the source-type that you'd rather the source use
    3) The user will respond with an answer to your question, using the chosen source. You can tap the message to look at why that model was chosen and the metadata behind the query (i.e. which chunks of the source were used to answer the question). 
  </Accordion>

  <Accordion title="Source mentioned">
    1. If you know which source should be used for a specific question, pass it in via the `@nickname` syntax
    2. The model will respond with the answer to your question, using the chosen source
    3. You can tap the message to look at why that model was chosen and the metadata behind the query (i.e. which chunks of the source were used to answer the question). 
  </Accordion>
</AccordionGroup>

## Rate your responses to improve responses overtime

Rating your responses lets Solari agents improve over time

1. Every user can give feedback on the answers received
   1. `thumbs up`  for helpful answers
   2. `thumbs down` for unhelpful answers
   3. Add a `note` about the answer for any qualitative feedback
2. Solari also tracks how often the user has to correct the model router to use a different source. This helps Solari get better at choosing the right source as your team uses the agent more

See how to view these analytics.