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

# Mastering chat for deep customer discovery

Go beyond keyword searching and start a conversation with your customer feedback. In this guide, you’ll learn how to use chat (beta) to get to the heart of your data, faster. We’ll cover how to craft the right questions to ask to uncover everything from granular details in a single interview to broad, cross-cutting themes across your entire workspace.

***

## Understand your analysis scope

To get the best results, you first need to know where you’re asking your question. Chat automatically adjusts its scope based on your location in Dovetail. The filter icon next to the chat bar tells you which level you’re on.

There are three scopes for your analysis:

<CardGroup cols={2}>
  <Card title="Data object" icon="file">
    * This is the most focused level, looking at a single piece of data like one
      interview transcript, a video, or a specific note. It's perfect for drilling
      down into the details of one person's experience.\ \ You can use it to ask
      "*What usability issues does this participant encounter?"*
  </Card>

  <Card title="Project" icon="sparkles">
    * This scope synthesizes information across all data within the project
      you’re currently viewing. It's ideal for understanding the combined findings
      of a single research study. - - - You can use this to ask "What feature
      requests appear most frequently in these customer interviews?" or "What
      visual design feedback appears in this usability testing?"
  </Card>

  <Card title="Workspace" icon="globe">
    * The broadest scope, this queries across all projects and data within a
      folder or your entire workspace. Use it to spot high-level patterns and
      trends across multiple studies. - - - You can use it to ask *"What
      onboarding problems do customers mention most often across all our
      projects?*
  </Card>
</CardGroup>

<video autoPlay muted loop playsInline className="w-full" src="https://videos.ctfassets.net/8fl1jrx919na/1rjjAIGYiBtSJfQjaj9mVk/46bf6b765f8370c65c1b859c69488211/Chat_in_project_and_workspace.mp4" />

<Tip>
  The broader your scope (from a single data object to an entire workspace), the
  more specific and targeted your question needs to be to get a clear answer.
</Tip>

***

## Craft effective questions with the specificity principle

The quality of your answer depends entirely on the quality of your question. Vague questions lead to vague results. The system works best when you use specific keywords that are likely to appear in your data.

<CardGroup cols={2}>
  <Card title="Instead of" icon="circle-x" horizontal>
    * *"What are common themes?"* -
  </Card>

  <Card title="Try" icon="circle-check" horizontal>
    * *"What did customers say about our mobile app performance?"*
  </Card>

  <Card title="Instead of" icon="circle-x" horizontal>
    * *"Any insights about pain points?"*
  </Card>

  <Card title="Try" icon="circle-check" horizontal>
    * *"What frustrations do users describe when trying to complete their
      profile setup?"*
  </Card>
</CardGroup>

The most effective questions use concrete terms, as well as specific parts of your product or service like "onboarding," "mobile app," or "profile setup".

<Tip>
  Review your results. How did the specificity of your question affect the
  answer? Don't forget to check the citations to see the original context of the
  quotes.
</Tip>

***

## Speak your customers' language

Frame your questions using the natural language your customers would use. This helps chat find the most relevant information directly from the source. Think about the exact words and phrases that might appear in your customer feedback. For example:

* "What problems do customers mention with billing?"
* "How do users describe the signup process?"
* "What complaints appear about our customer support?"
* "What do customers like about our dashboard?"

***

## Start simple, then dig deeper

Analysis is a conversation. You don’t have to find everything in a single question. Ask follow-up questions, probe for more nuance, challenge an assertion, or pivot your line of inquiry entirely. It’s like having a real conversation with your data, allowing you to uncover insights that might have remained hidden.

Start with a broad, yet specific, query and then use follow-up questions to drill down into the details. For example:

<Steps>
  <Step title="Initial question">
    "What challenges do customers face with our billing process?"
  </Step>

  <Step title="Follow up">
    "What specific billing errors do customers mention?"
  </Step>

  <Step title="Deep dive">
    "How do customers describe the impact of these billing issues on their
    business?"
  </Step>
</Steps>

<Tip>
  Enable chat for your workspace to ask questions about your customer feedback data. Start by opening a single interview or document in a project and click Chat to discover actionable takeaways.

  Don't forget to check the citations to see the original context of any quotes cited.
</Tip>