Skip to main content

Report Chat

Explore your data conversationally with AI-powered analysis chat

Report Chat Interface

Chat with your research data using natural language. Ask questions, explore findings, and uncover insights beyond the standard report.

Accessing Chat

  1. Navigate to Analysis tab

  2. Click Chat or Assistant

  3. Start asking questions

Chat Panel Layout

Area

Purpose

Message history

Previous questions and answers

Input field

Type your question

Suggestions

Common question starters

Citations

Links to source responses

Chat List Panel

View and manage multiple conversations.

Creating Chats

Click New Chat to start fresh. Each chat maintains its own history and context.

Switching Chats

Click any chat in the list to resume that conversation.

Chat Organization

  • Chats auto-title based on first question

  • Recent chats appear at top

  • Search across all chats

Voice Transcript Input

Use voice to ask questions:

  1. Click microphone icon

  2. Speak your question

  3. Transcript appears in input

  4. Send or edit before sending

Chat Streaming

Responses stream in real-time:

  • See answers as they generate

  • Don't wait for complete response

  • Cancel if going wrong direction

Report Chat Provider

The chat understands your specific data:

  • All survey responses

  • Generated personas

  • Identified themes

  • Quantitative results

Data Access

Chat can reference:

  • Individual responses

  • Aggregate statistics

  • Cross-tabulations

  • Qualitative themes

Creating Report Chats

New Conversation

Each new chat starts fresh with:

  • Full access to study data

  • No prior context

  • Clean slate for exploration

Continuing Conversations

Return to previous chats to:

  • Follow up on findings

  • Ask related questions

  • Build on prior analysis

Scoped Chat Layout

Chat is scoped to your current report:

  • Questions answered from this study only

  • Personas and themes available

  • Report context understood

Chat Suggestions

Starter Questions

Common starting points:

  • "What are the key findings?"

  • "Summarize the main themes"

  • "What surprised you about the data?"

Follow-Up Suggestions

After each response:

  • Related questions appear

  • Deeper exploration options

  • Alternative angles

Question Types to Ask

Exploratory

Understand broad patterns:

  • "What are the main reasons for dissatisfaction?"

  • "What themes emerge about pricing?"

  • "How do personas differ?"

Specific

Get targeted answers:

  • "How many mentioned sustainability?"

  • "What do people say about the mobile app?"

  • "How do Panel A and B compare?"

Comparative

Understand differences:

  • "How do satisfied vs. dissatisfied differ?"

  • "What distinguishes heavy users from casual?"

  • "How does perception vary by age?"

Summary

Get concise overviews:

  • "Summarize concerns about the new feature"

  • "What are the top 3 recommendations?"

  • "Quick overview of brand perception"

Effective Questions

Be Specific

Less Effective

More Effective

"What do people think?"

"What do people think about our customer service?"

"Tell me about the data"

"What are the main themes in open-ended responses?"

Ask One Thing at a Time

Break complex questions into steps:

  1. "What do people think about the price?"

  2. "How does this compare to quality perceptions?"

  3. "What do people say about competitors?"

Understanding Responses

Response Structure

Answers typically include:

  • Direct answer: Main finding

  • Evidence: Supporting data/quotes

  • Context: Caveats or nuances

  • Citations: Links to sources

Following Up

After each response:

  • Ask for more detail

  • Request examples

  • Explore related topics

  • Verify findings

Best Practices

πŸ’‘ Tip: Start broad, then narrow. Begin with general questions, follow up on interesting areas.

πŸ’‘ Tip: Verify important claims. Click through to original responses.

πŸ’‘ Tip: Ask for examples. Request specific quotes to understand perspective.

πŸ’‘ Tip: Explore contradictions. Ask about conflicting responses or minority views.

Did this answer your question?