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

# Tips for good results

> A few habits that get you better answers from Token.

Token works best when you treat it like a capable teammate who is new to your
world. A little context goes a long way.

<CardGroup cols={2}>
  <Card title="Be specific" icon="bullseye">
    Say what you want and what a good result looks like. "Summarize this thread in
    three bullets for the team" beats "summarize this."
  </Card>

  <Card title="Give context" icon="circle-info">
    Point Token at the doc, the link, or the channel it needs. It only knows what
    it can reach.
  </Card>

  <Card title="One task at a time" icon="list-ol">
    For multi-step work, let it finish one step and confirm before piling on the
    next.
  </Card>

  <Card title="Correct it" icon="pen">
    If it gets something wrong, just say so in plain language. Token adjusts and
    tries again.
  </Card>
</CardGroup>

## When Token asks for approval

That pause is normal and healthy. It means Token is about to do something worth a
look. Approve it, decline it, or tell Token to take a different approach. See
[Approvals](/permissions).

## If you are not getting what you want

* **Check the tool is set up.** Token can only act through tools
  [shared with that chat](/integrations).
* **Mention it in group channels.** In a shared channel, Token only responds when
  you [get its attention](/reaching-token).
* **Break it down.** A big, vague ask often works better as a couple of clear,
  smaller ones.