Skip to content
scrutinizr.
Platform overviewAnalysis & machine learning

Find the differences. Understand the patterns.

Check whether sources agree, see what changed between snapshots, or look for unusual groups in your data. Guided analysis helps you choose a method and return to the passages or rows behind the result.

Explore workspaceSee what you can do
Illustrative product journey

Where should we look more closely?

  1. Evidence

    A website promise, an internal policy and order records.

  2. Analysis

    Compare the statements; check numeric fields for unusual values.

  3. Next check

    Inspect the source excerpts and flagged orders before deciding what changed.

WHAT YOU CAN DO

The tools, with a clear purpose.

Compare claims

Review different statements across websites, documents and feedback. Open the cited excerpts and see which checks were used.

Explain record changes

Compare imported snapshots using a match key. Inspect added, removed and changed records alongside missing or repeated values.

Discover groups

Choose numeric fields from a table snapshot. Clustering groups similar records so you can inspect what they share.

Find unusual values

Run outlier checks to locate records that stand apart. Review the original rows before treating a signal as a problem.

Explore relationships

See which numeric fields move together. Use the result to form a question to investigate; a relationship does not establish a cause.

Use optional models

Built-in numeric analysis works without creating a model. Configured AI interpretation and a compatible Azure Machine Learning endpoint can extend the workflow.

HOW IT FLOWS

A path through the work.

  1. 01

    Choose the evidence

    Link the sources your product question needs. For numeric analysis, select a table snapshot and the fields to examine.

  2. 02

    Run a guided check

    Choose a comparison or pattern method. Review its coverage, method and limitations alongside the result.

  3. 03

    Inspect and follow up

    Return to the cited passage or row. Save a review, build a report or create a task for the next check.

Azure Machine Learning is an optional connection to an existing compatible endpoint. You do not need an Azure workspace to use the built-in pattern tools.

CONNECTED COMPONENTS
Website researchDocuments & dataFeedback & researchProduct journeys & outcomes

Start with the question you need to answer.

Explore the sample workspace or bring your own evidence.

Start freeExplore workspace