Generating Research Reports

Koji's research reports pull together findings from all the qualifying interviews in a study into a single, structured document. Instead of reading every transcript individually, you get an executive summary, key takeaways, theme analysis, traceable charts, written findings, statistics, and actionable recommendations — ready to share with your team.

Quality Filter: Only Substantive Data

An important detail: reports only include interviews with a quality score of 3 or above. This quality filter ensures your aggregate analysis is built from substantive data rather than being diluted by low-quality responses. The results page shows a report-eligible count so you always know how many interviews will feed into your report.

Learn more about quality filtering in How the Quality Gate Works.

What's in a Research Report

Each generated report includes several sections designed to give you a complete picture of your research findings:

Executive Summary

A concise overview of the most important findings from your study. The executive summary distills your qualifying interviews into key takeaways that anyone can understand in two minutes. It's written for stakeholders who need the bottom line without wading through details.

Key Takeaways

Critical, high, and medium priority findings extracted from the data. Each takeaway is a specific, actionable insight ranked by importance and backed by evidence from the interviews.

Theme Analysis

The report identifies the most prominent themes across all qualifying interviews, with frequency data and traceable citations. Each theme includes:

• Description: What the theme represents and why it matters
• Frequency: How many interviews touched on this theme
• Supporting quotes: Direct participant quotes that illustrate the theme, with citation links back to the source interview

Themes in the report aren't just a list of topics — they're synthesized findings that connect what multiple participants said into coherent narratives. Learn more in Understanding Themes & Patterns.

Charts and Visualizations

Reports include several types of traceable charts, each linked back to the source interviews:

• Pie charts: Sentiment distribution across interviews (positive, neutral, negative, mixed)
• Horizontal bar charts: Pain point frequency, feature request frequency, positive highlight frequency
• Vertical bar charts: Option frequency for choice questions
• Distribution charts: Quality score distribution, scale response distributions with mean, median, and mode
• Quote citations: Notable quotes with attribution to specific interviews

Structured Question Charts

If your study uses structured questions, reports produce rich quantitative visualizations:

• Scale questions produce distribution charts showing how participants rated each item, with calculated mean, median, and mode. For 0-10 scales, Koji automatically calculates an NPS (Net Promoter Score), categorizing respondents into promoters, passives, and detractors.
• Single and multiple choice questions produce bar charts showing option frequency and percentages — how many participants selected each option.
• Ranking questions produce average position charts showing where each item landed in participants' rankings.
• Yes/No questions produce pie charts showing the binary distribution of responses.

Every chart includes traceable citations linking back to the source interviews, so stakeholders can click through to verify any data point. This gives your team the quantitative charts they need, backed by qualitative context from the AI conversation.

Written Findings

A detailed narrative section that expands on the key takeaways with full context, evidence, and analysis. Written findings provide the depth that executives skip but product managers and researchers rely on.

Recommendations

Based on the patterns found across interviews, the report suggests concrete actions you could take. These recommendations are grounded in participant data and connected to specific themes and quotes, so you can trace each suggestion back to its evidence.

Stat Cards

Summary statistics presented as visual cards, giving you an at-a-glance overview of key metrics from the study.

Question Coverage

A breakdown of how well each research question from your study brief was covered across interviews. This helps you identify gaps — if a particular question was barely addressed, you may need more interviews or a revised approach.

How to Generate a Report

1. Complete enough qualifying interviews Reports work best with sufficient data. While you can generate a report with just a few interviews, the analysis becomes more robust with more participants. Most researchers find that meaningful patterns emerge after 5-8 quality-qualifying interviews (scoring 3+).

2. Navigate to your study's report section Open your study from the dashboard and look for the Report tab or Generate Report button.

3. Click Generate Report Koji's AI will analyze all qualifying interviews in the study, identify cross-cutting themes, select the best supporting quotes, calculate statistics, generate charts, and produce a structured report. This typically takes 30 to 60 seconds depending on the number of interviews.

4. Review the generated report Once complete, the report appears with all sections described above. Take time to read through it and cross-reference findings with individual transcripts where needed.

Report Versioning

Each time you generate a report, Koji creates a new version. This is important for several reasons:

• Progressive research: As new interviews come in, you can generate an updated report that incorporates the latest data. The previous version is preserved with its version number.
• Point-in-time snapshots: Each version captures the state of your research at a specific moment. A version marked as "current" (via the isCurrent flag) is the latest one.
• No data loss: Generating a new report never overwrites or deletes a previous version. You always have access to your full report history.

Plan Access

Research reports are available on all plans. Credits are the only gate — generating or refreshing a report costs credits from your balance. Check the Plan Comparison Guide for credit costs and allocations across plans.

Tips for Better Reports

• Wait for enough qualifying data: While you can generate a report after just a few interviews, waiting until you have at least 5-8 interviews scoring 3+ produces more reliable and convincing results. Theme patterns need repetition to be meaningful.

• Add structured questions for richer charts: Studies with structured questions produce reports with quantitative visualizations — scale distributions, choice breakdowns, ranking charts — that give stakeholders the data-driven evidence they expect.

• Review your study brief before generating: Make sure your research objectives are clearly defined in the study brief. The report's analysis is anchored to those objectives, so a clear brief produces a more focused report.

• Use reports iteratively: Generate a report mid-study to check for emerging patterns. Use those early findings to refine your approach, then generate a final report when all interviews are complete.

• Combine with individual insights: Reports give you the big picture. For specific details, always refer back to individual AI-generated insights and transcripts.

• Share thoughtfully: Reports are designed for stakeholders, but add your own context when presenting. You know the business situation, competitive landscape, and strategic priorities better than any AI. See Publishing & Sharing Reports for sharing options.

Key Things to Know

• Reports filter to quality 3+: Only interviews scoring 3 or above on the quality scale are included in report analysis. This ensures your findings are built from substantive data.
• Report generation takes a moment: Complex studies with many interviews may take 30-60 seconds to process.
• Reports complement, not replace, individual analysis: The aggregate view is powerful, but always dig into individual transcripts for nuance and context.
• Traceable citations: Every theme, quote, and chart data point links back to its source interview, enabling full traceability.

Related Articles

• AI-Generated Insights — Per-interview analysis that feeds into aggregate reports
• Understanding Themes & Patterns — How recurring themes are identified across interviews
• Understanding Quality Scores — How quality filtering shapes your reports
• How the Quality Gate Works — The quality threshold that protects your data
• Plan Comparison Guide — Compare plans to see credit allocations
• Publishing & Sharing Reports — Make your reports accessible to stakeholders
• Structured Questions Guide — Design questions that produce rich report visualizations

Frequently Asked Questions

Q: Can I generate a report with just one or two interviews? A: Technically yes, but the report will have limited value. With only one or two data points, there aren't enough interviews to identify meaningful cross-cutting themes. We recommend at least five qualifying interviews for a useful report.

Q: Does generating a new report delete the previous one? A: No. Each generation creates a new version. All previous versions remain accessible, giving you a history of how your findings evolved over time.

Q: Why are some interviews excluded from my report? A: Reports only include interviews with a quality score of 3 or above. Check the report-eligible count on your results page to see how many interviews qualify. Low-scoring interviews still have individual insights but are excluded from aggregate analysis.

Q: Can I customize what appears in the report? A: Reports are generated based on your study's research objectives and all qualifying interviews. The AI determines the most relevant themes, quotes, and recommendations based on the data.

Q: Can I export or share my report? A: Yes, reports can be published with a shareable public URL. See Publishing & Sharing Reports for details.

Further reading on the blog

• Agile User Research: How to Run Continuous Research in Sprint Cycles (2026) — Most teams know they should do user research every sprint. Almost none actually do. Here's the practical playbook for integrating continuous
• AI Agents for User Research in 2026: How Autonomous Research Is Reshaping Customer Insight — AI agents are taking over user research in 2026 — moderating interviews, synthesizing themes, and producing insight reports in hours. The fu
• Best AI Market Research Tools in 2026: The Complete Buyer's Guide — AI has fundamentally changed market research. This guide compares the leading AI market research platforms—from AI-native interview tools li

<!-- further-reading:blog -->