Insights Chat: Ask Any Question About Your Research Data with AI

The Insights Chat is an AI-powered query interface built into Koji that lets you ask natural-language questions about your collected research data. Instead of re-reading every transcript or waiting for a full report refresh, you can type "What did B2B customers say about pricing?" and get an instant, evidence-backed answer with supporting quotes.

Think of it as a research analyst who has read every transcript in your study and is always available to answer follow-up questions.

How Insights Chat Works

When you generate a report for your study, Koji processes all completed interviews, extracts themes, maps answers to your structured questions, and builds an indexed knowledge base from your transcripts. The Insights Chat draws on this knowledge base to answer your questions.

Unlike a simple keyword search, the Chat understands context and intent. It can:

• Synthesize patterns across multiple interviews: "What were the most common reasons for churn?"
• Surface specific quotes to support a claim: "Show me quotes about onboarding confusion"
• Compare responses across segments: "How did enterprise customers describe pricing versus SMB customers?"
• Answer quantitative questions from structured data: "What percentage of participants rated satisfaction 8 or above?"
• Generate stakeholder-ready summaries: "Give me a two-paragraph executive summary of the key findings"

The Chat is grounded in your actual research data. Every answer it provides is traceable back to specific interviews and quotes — not hallucinated or generalized from outside sources.

Accessing Insights Chat

Insights Chat is available in the Insights tab of any study that has completed interviews and a generated report.

Requirements:

• At least 1 completed interview that passed the quality gate
• A generated report (or refreshed report for studies with recent completions)
• Insights plan or higher

Look for the chat input at the bottom of the Insights Dashboard. Type your question and press Enter.

What to Ask: Example Queries by Category

Theme Discovery

• "What were the top 3 pain points participants mentioned?"
• "What themes came up most frequently across all interviews?"
• "Were there any surprising or unexpected findings?"
• "What topics did participants bring up that I did not ask about?"

Quote Retrieval

• "Show me quotes about the onboarding experience"
• "What did participants say about our pricing?"
• "Find me the most compelling quote that illustrates the main friction point"
• "Show me a quote I could use in a stakeholder presentation"

Segment Analysis (requires participant attributes from personalized links)

• "Did enterprise users have different concerns than small business users?"
• "How did responses differ between users who completed onboarding and those who did not?"
• "Which participant segment expressed the most urgency around this problem?"

Structured Data Queries (requires structured questions in the study)

• "What was the average satisfaction score?"
• "What percentage of participants answered yes to the referral question?"
• "Which feature option was ranked highest overall?"
• "Show me the distribution of scores on the ease-of-use question"

Hypothesis Testing

• "Did participants confirm the hypothesis that checkout flow is confusing?"
• "Is there evidence that pricing is a barrier to expansion?"
• "How strong is the signal on the feature request theme?"

Stakeholder Preparation

• "Summarize the three most important findings in two sentences each"
• "Give me a one-paragraph executive summary"
• "What are the top three recommendations based on the data?"
• "What questions might stakeholders ask, and what do the data say?"

Insights Chat vs. Research Reports: When to Use Each

Research reports are structured, visual documents designed for sharing — they display distributions, aggregate themes, and selected quotes in a formatted layout optimized for presentations and stakeholder reviews.

Insights Chat is conversational and exploratory — it is for the researcher, not the audience. Use it to:

• Investigate a hunch before adding it to a formal report — is this a pattern or an outlier?
• Answer a stakeholder question on the fly — even during a live presentation
• Dig into a segment or edge case that the structured report does not highlight
• Cross-reference findings from multiple angles without generating a new report

The two tools complement each other. Use the report for communication; use the Chat for investigation and discovery.

Structured Questions Make Insights Chat Dramatically More Powerful

When your study includes structured questions, the Insights Chat can answer quantitative questions with precision instead of estimation.

Koji supports six structured question types:

Combining quantitative precision with qualitative depth is what makes AI-native research fundamentally different from traditional survey tools. You get the statistical clarity of a structured questionnaire and the nuanced texture of an in-depth interview — in one study, queryable in seconds.

Practical Workflow: Using Chat During a Stakeholder Review

You are presenting findings to your product leadership team. The CEO asks a question you did not specifically prepare for: "What did early-stage customers say about the API documentation specifically?"

Without Insights Chat, you would say "I will follow up on that." With Insights Chat open on a second screen, you type the question and get an answer grounded in your actual research within seconds — with supporting quotes you can read aloud.

This is the difference between research as a static document and research as a living, queryable knowledge base.

Practical Workflow: Competitive Intelligence from Customer Interviews

You have run 40 customer discovery interviews. You want to understand the competitive landscape from your customers' perspective. Ask the Chat: "Which competitors did participants mention, and what did they say about them?"

The Chat synthesizes competitive mentions across all 40 transcripts, giving you a clear picture without manually reviewing each transcript. What would have been a 2-hour analysis task becomes a 30-second query.

Best Practices

Ask specific questions. "What did people say about pricing?" works, but "What specific concerns did participants raise about annual contract pricing?" works better.

Reference your research goals. The Chat knows your study brief. Leverage that: "Given our goal of understanding churn drivers, what is the most important finding in this data?"

Iterate conversationally. If the first answer is too broad, drill down: "You mentioned support friction — give me quotes specifically about response times rather than resolution quality."

Refresh your report before querying new data. If you have added interviews since your last report generation, run a report refresh so the Chat has access to the latest data.

Use Chat for segment exploration, reports for segment presentation. Chat is faster for discovery; reports are better for structured visual comparisons.

What Insights Chat Does Not Do

Insights Chat is a research analysis tool, not a general AI assistant. It will not:

• Generate insights about topics not covered in your research data
• Make statistical claims from samples smaller than 5 interviews (it will flag low sample sizes)
• Replace the judgment of a trained researcher about research validity
• Search the web or draw on information outside your study

Every answer is grounded in your collected interviews. If a topic was not discussed in your research, the Chat will tell you rather than fabricate an answer.

Availability and Plans

Insights Chat is available on the Insights plan and above. Free plan users can view the basic report summary but cannot use the conversational query interface.

On the Interviews plan, Insights Chat is available with no session limits.

Related Resources

• Structured Questions Guide — Add quantitative question types that make Chat queries more precise and powerful
• Generating Research Reports — The underlying data layer that powers Insights Chat
• AI-Generated Insights — How Koji automatically surfaces themes and patterns from your interviews
• Understanding Themes and Patterns — What the AI extracts from your interview data
• Turning Interviews Into Insights — From raw transcripts to actionable conclusions
• Insights Dashboard — The complete Insights feature set overview