Research · Analysis

It does not estimate
the number.

It computes it, and shows what it ran. Cross-tabs with significance, custom table definitions, and named statistical functions — with an AI layer that chooses and runs them rather than guessing at an answer.

An AI that describes statistics is not doing statistics. The functions here are real, and the chat calls them.

The analysis

Functions, not adjectives.

Named functionsThe statistical operations available, invoked explicitly, so what was run is knowable.
Custom table insightDefine the table you want rather than choosing from a menu of shapes.
QueryAsk the dataset directly for the cut you need.
The AI layer

Chooses and runs.

AI chat over the dataAsks what you meant, picks the function, runs it, and answers with the result.
Insights as recordsFindings kept rather than regenerated, so a deck can cite something stable.
Sharing it

Embeds and dashboards.

Embeddable outputsSo a client sees the dashboard rather than a screenshot of it.
Per projectFindings organised against the project they belong to.
Sharing findings

Shareable, and measured.

Share tokensA finding shared by link, with the views against it counted — so you know whether the client opened it.
Regenerate and replayAn insight re-run against updated data rather than rebuilt, which is what makes a tracker’s findings comparable.
Embeds with their own keysA dashboard embedded in a client portal, with a key you can rotate and analytics on its use.
Nearby

Related capabilities.

Questionnaires · Open-end coding · Data files

The number, and what produced it.

Because it ran a function rather than forming an impression.

Get started → See pricing