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

# Ask Pillow about your data

> Ask questions about your interviews and get answers grounded in real responses, with clickable sources.

Ask Pillow is built into every study. Ask a question about your interview data in plain language and get an answer grounded in real participant responses, with clickable sources you can trace back to the original transcript.

<Info>
  Ask Pillow was previously called Research Chat. The panel and features are the same.
</Info>

## Open Ask Pillow

Open any study from your dashboard. The Ask Pillow panel sits on the right side of the page. It's always there, no navigation required. The first time you open a study, a hint points you to it above the analysis tabs.

## Ask a question

Type any question about your study:

* "What are users saying about pricing?"
* "What frustrates users about checkout?"
* "How do users feel about onboarding?"

Pillow answers from your actual interviews. Every answer includes **source references**. Click a source to jump to the original transcript.

<Tip>
  Specific questions ("What frustrates users about checkout?") produce richer answers than broad ones ("Tell me about my users").
</Tip>

## Use quick prompts

When you open the panel, quick prompts help you start without typing:

* **Summarize key findings**: an overview of what your interviews reveal
* **Executive summary**: a high-level narrative for stakeholders
* **Most requested features**: what users want added or improved
* **Interviews we should read**: the most valuable interviews for your team
* **Most frequent issues**: recurring problems across interviews

Click a prompt to run it.

## Manage threads

Use tabs to explore several angles at once:

* **New thread**: click **+** to start a fresh thread
* **Switch threads**: click any tab
* **Close a thread**: click the **x** on a tab
* **Reopen past threads**: use the history dropdown

Each thread keeps its own history. For example, you might run one thread for product feedback and another for UX issues.

## Add context

Focus an answer on a specific slice of your data by attaching context to your next question:

* **From an interview**: click **Add to chat** in the interview header to load that interview
* **From the Overview tab**: click **Add to chat** on a question group to dig into one pattern

Context appears as a pill above the input. Click the **x** on the pill to remove it before sending.

## Read visual evidence

Answers can include charts inline: bar and pie charts that appear automatically when an answer is backed by a specific pattern, so you see the evidence alongside the text.

## Filter by metadata

Scope answers to a specific audience segment. Click the filter icon in the input and select metadata values:

* **Locale**: filter by participant language
* **Country**: filter by location
* **Device**: filter by desktop, mobile, or tablet
* **Custom fields**: filter by any [metadata](/guides/customize-metadata) you passed in

While a filter is active, every answer is scoped to matching interviews. Remove the filter to return to the full dataset.

## Share answers

Use **Share** to send teammates exactly what you want them to see:

* **Share a thread**: click **Share** at the top of the panel
* **Share one answer**: click **Share** on a specific message

When you create a link, choose:

* **Visibility**: Organization-only (requires sign-in) or Public (anyone with the link)
* **Expiration**: Never, 7 days, or 30 days

## Overview

The **Overview** tab surfaces patterns across your interviews automatically: the questions your agent asked most and how participants answered.

Each group shows:

* **Question**: the grouped question theme
* **Ask count**: how often the agent asked it
* **Answer buckets**: how participants responded, as percentages

Look for questions with widely varied answers; they often reveal segments with different needs.

<Info>
  Overview becomes available once a study reaches **100 interviews**. When new interviews arrive, a banner shows how many are available, so click **Regenerate** to update. Before then, you can still ask Pillow about the interviews collected so far.
</Info>

## Small-sample note

When a study has fewer than **10 interviews with replies**, Ask Pillow flags that the sample is still small.

## Tips for sharper analysis

<Steps>
  <Step title="Start broad, then narrow">
    Run a quick prompt for the high-level view, then ask specific follow-ups.
  </Step>

  <Step title="Split topics across threads">
    Keep one thread per theme so each stays focused.
  </Step>

  <Step title="Add context to go deep">
    Attach an interview or an Overview group when you want the answer to focus on one slice.
  </Step>

  <Step title="Share the executive summary">
    Send stakeholders the high-level narrative instead of the raw transcripts.
  </Step>
</Steps>

## What's next

<CardGroup cols={2}>
  <Card title="Review interviews" icon="eye" href="/guides/review-interviews">
    Read individual interviews in detail.
  </Card>

  <Card title="How insights work" icon="lightbulb" href="/concepts/insights">
    Understand Overview and Ask Pillow.
  </Card>

  <Card title="Track issues" icon="list-checks" href="/guides/issue-tracking">
    See bugs and requests clustered across interviews.
  </Card>

  <Card title="Broadcast your study" icon="share" href="/guides/broadcast-study">
    Collect more interviews to strengthen your analysis.
  </Card>
</CardGroup>