AI Analyst

Answers you can check, not just a summary.

The AI Analyst runs the statistics on your responses, finds what drives your outcome, splits results by any question and writes recommendations that point to the numbers behind them.

Try it on Professional Take the product tour

Professional free for 7 days. Cancel anytime.

  • Key drivers
  • Significance tests
  • Compare by any question
  • Evidence on every finding
An AI Analyst report with findings and evidence
Sample data

How it works

From the first response to a decision

Quick reads while answers come in, and a deep analysis when you need to decide what to do.

AI summary

A plain-language summary of your results

On the results page, Clue writes a short summary of what respondents said: the headline numbers, the main themes in open answers and what stands out.

  • Written from your real responses
  • Open-text themes in a few lines
  • Uses one AI credit
Sample data

Instant intelligence

Health and churn risk at a glance

Survey intelligence scores the overall health of a survey from satisfaction, loyalty, engagement and sentiment, flags respondents at risk of churning and lists the actions to take first.

  • Health score with a short explanation
  • At-risk respondents and churn summary
  • Re-run it with the latest responses
The AI Analyst executive summary with a health score
Sample data

Deep analysis

A full analysis run, step by step

Pick an outcome and a goal, and the analyst works through descriptive statistics, significance tests, probability estimates, segments and open-text themes, then checks data quality before it writes the report.

  • Five specialist steps plus quality checks
  • Statistics computed by a stats engine, not guessed
  • Written in the language you use Claform in
Sample data

Key drivers and tests

What actually moves your score

Key-driver analysis shows which questions go with a higher or lower outcome and how much of the variation the model explains. Chi-square, t-tests, ANOVA and correlations are corrected for multiple testing.

  • Effect sizes and 95% intervals
  • Importance versus performance chart
  • Associations are labelled as associations, not causes
AI Analyst key driversSample data

Compare by

Split every measure by one question

Choose a question, such as location or plan, and every measure in the report is compared across its answers. Small groups are labelled, and groups under ten responses are not shown.

  • Averages, NPS and answer shares per group
  • The clearest difference called out
  • Uses no credits
Sample data

Recommendations

Recommendations that cite their evidence

Each recommendation links to the findings it is based on. Select a citation to see the statistic behind it. A recommendation that cites nothing real is dropped before you see it.

  • Evidence chips with test, effect size and sample size
  • Priority and timeline for each action
  • Unverified items are marked as such
AI Analyst recommendationsSample data

Built for trust

The details that make the report usable

Know how sure the analyst is, run it on a schedule and share the result.

  • Confidence breakdown

    The headline confidence is built from four signals: sample size, how many tests are significant, effect sizes and how many recommendations were verified. Small samples are capped.

  • Scheduled analyses

    Run the analyst daily, weekly or monthly at a time and time zone you choose. If your credits run out, the run is skipped and you are told why.

  • PDF report

    Download the report as a PDF, with your team branding when you have white-label set up.

  • Shared with your team

    Teammates who can open the survey see the same reports, so nobody runs the same analysis twice.

  • Data-quality caveats

    An integrity step flags small samples, unreliable tests and other caveats, so a weak result is never presented as a strong one.

  • Credits you can predict

    A deep analysis run uses 3 AI credits. Comparing by a question is free. If a run fails, its credits are returned.

Every plan includes a monthly allowance of AI credits; the AI Analyst is available on Professional and higher. See pricing for each plan.

The AI Analyst in numbers

Credits per deep run
3Credits per deep runReturned if the run fails
Credits to compare by
0Credits to compare bySplit a finished report by any question
Confidence signals
4Confidence signalsSample, significance, effect size, evidence
Schedule options
3Schedule optionsDaily, weekly or monthly

AI Analyst questions

Does the AI make up the numbers?

No. The statistics are computed by a deterministic stats engine on your responses. The AI writes the narrative and the recommendations, and each recommendation must cite real findings; ones that cite nothing real are removed.

How many responses do I need?

You can run it on any survey with completed responses, but small samples limit what the tests can show. The report flags small samples and caps its confidence when the data is thin.

Which plan includes the AI Analyst?

The AI Analyst is available on Professional and higher. Every plan has a monthly AI credit allowance; a deep analysis run uses 3 credits and comparing by a question uses none.

Can I run it automatically?

Yes. Schedule it daily, weekly or monthly for a survey. Each scheduled run uses 3 credits and is skipped, with a notification, if your credits are used up.

Works with

Let the analyst read your results.

Start on the Free plan, then try the AI Analyst on Professional with the trial.

Try it on Professional Take the product tour