Manual analysis
- Hours reading individual comments one by one
- Patterns buried in noise, hard to spot recurring themes
- No structured way to share findings with your team
- Feedback piles up and never gets analyzed
AI-powered analytics
Stop reading every comment
AI Reports creates a structured first pass over your feedback: an executive summary, suggested themes with user quotes, and possible fixes for your team to review against the underlying comments.
Available on the Unlimited plan. Generate reports on demand.
Sample AI Report
Illustrative data
Volume
312 entries
Executive summary
Sentiment is +12% this week. Auth examples drove the most negative feedback, accounting for 23 of 312 entries.
Top recurring theme
Auth examples missing refresh-token logic
-68%“The token example doesn't show how to refresh it. We had to read the source code.” — 23 mentions
Recommended fix
Add a refresh-token example to /api/auth/tokens. The most mentioned issue this period, with 23 related comments.
Generated in3.2s
Learn how reports workAI Reports analyzes feedback collected across your documentation sites
You collect feedback.
Then what?
Reading hundreds of comments to spot a pattern is the kind of work that fills a Wednesday and gets postponed to next Wednesday. Then it piles up. Then nobody reads any of it. AI Reports does it in seconds, any time you ask. The five feedback metrics worth tracking →
With AI Reports
- One click analyzes all your feedback in seconds
- AI surfaces recurring themes with mention counts
- Shareable report with executive summary built in
- Concrete recommendations you can act on today
How it works
Three steps. One workflow.
Collect, track, and analyze user feedback in one tool. No exporting CSVs to ChatGPT, no copy-pasting between dashboards.
- 01
Collect with the widget
Drop a feedback widget on any page. Users submit ratings, comments, and screenshots in seconds.
Was this helpful? - 02
Track in the dashboard
All feedback lands in one inbox. Filter, tag, and search across pages, products, or time ranges.
Auth example unclear23Mobile screenshots stale18New tutorial works great+9 - 03
Analyze with AI
One click. AI reads everything, surfaces themes, ranks priorities, and writes the report.
Analyzing 312 entries...✓ 9 themes · 5 fixes
What's in an AI ReportFrom high-level overview
From high-level overview
to specific actions.
Every report is structured the same way. Open it, skim the summary, drop into themes, ship the recommendations. Send it to your team in Slack as-is.
Section 01
Executive Summary
Feedback volume, overall sentiment, and the single most notable finding. Perfect for a quick status check or sharing with stakeholders who don't have time to read the full report.
Volume
312
feedback entries
Sentiment
+12%
vs last week
Themes
9
recurring patterns
Key finding
Auth examples drove the most negative feedback this week, accounting for 23 of 312 entries. Sentiment improved overall thanks to the new tutorial section, but the auth section remains a clear blocker.
Section 02
Most Requested Improvements
Up to 5 recurring themes extracted from your feedback. Each comes with a description, mention count, sentiment label, and a representative user quote pulled directly from the data.
P1
-68%Auth examples missing refresh-token logic
23 mentions across /api/auth/*
“The token example doesn’t show how to refresh it. We had to read the source code.”
P1
-54%Mobile screenshots out of date
18 mentions across /guides/*
“The screenshots don’t match the new dashboard at all. Confusing.”
P2
-41%Webhook setup unclear after step 4
12 mentions on /api/webhooks
“I followed steps 1-3 fine, then step 4 references a setting I can’t find.”
+ 2 more themes in the full report
Section 03
Actionable Recommendations
3 to 5 concrete suggestions for improving your content based on the feedback patterns. Not vague advice like "improve onboarding". Specific steps you can ship this week.
1
Add a refresh-token example to /api/auth/tokens
Single doc change addresses the most mentioned theme, with 23 related comments.
2
Re-shoot 6 mobile screenshots in /guides/installation
Last refresh was 14 months ago. Resolves 18 mentions.
3
Rewrite step 4-6 of webhook setup
Most users dropped feedback at step 5. Add a screenshot of the settings location.
Built for any team that collects feedback
AI Reports adapts to whatever you ship.
Documentation teams
Know which pages to rewrite first based on what confuses readers most. Stop updating docs based on gut feeling.
Product teams
Spot UX issues users keep reporting and prioritize fixes by user impact, not by who's loudest in the standup.
Education
Find which lessons confuse students most and improve learning outcomes before the next cohort hits the same wall.
Internal teams
Summarize QA feedback for stakeholders with a report you can share in seconds. No formatting, no slide decks.
Frequently asked questions
You can, once. Then you have to do it again next week, and the week after. AI Reports runs on the live feedback in your dashboard, with prompt scaffolding tuned for product feedback specifically. The output is structured (themes, mention counts, sentiment, recommendations) instead of a wall of prose. Plus, your feedback grows over time, and re-pasting CSVs into ChatGPT every Monday is the kind of process nobody actually keeps doing.
Usually a few seconds for up to a few hundred feedback entries. Larger volumes take a bit longer but rarely more than a minute. Run a report on demand whenever you need one, and past reports stay saved in your dashboard.
AI Reports is available on the Unlimited plan with no feedback cap. Free and lower-tier plans can collect feedback in the dashboard, and you can upgrade to unlock AI Reports any time.
No. Your feedback is processed to generate your report and is not used for any model training. We don't share your data with third parties beyond the LLM provider that runs the analysis, and that provider's terms prevent training on your inputs.
Yes. Download the report as a PDF or copy the full text with one click. Most teams drop the summary into a #feedback channel for the whole company to see.
Yes. Feedback in any language is analyzed without any setup. The report itself is generated in English, with user quotes preserved in their original language.
Let AI analyze your feedback. Get back to building.
Generate a structured first pass over your feedback, then review the themes and suggested actions before deciding what to fix.
- No feedback limit
- Available on Unlimited plan
- Results in seconds
- Cancel anytime