The first output that is seen after you start an analysis report is the AI "So What" Executive Summary for the report. It provides an overview and links to the most important insights that matter so you can dig deeper within the Google Sheet.

The AI behind the summaries have been thoroughly and consistently tested for accuracy and usefulness. In the summary there are links to the insights that link back to the real calculations and source data to avoid hallucinations.

"So What" Insights are the ranked insights from AI that point to the real metrics and calculations that matter. The amount insights are controlled by you and ranked based on the goal you have for the report.

Unseen connections in the real metrics are referenced within the insights in a way that is easily understood. You can instantly see the calculations behind the insight, as well.

A key aspect of the "AI-guided exploration" are these insight deep links that guard against AI hallucinations. The deep links provide a connective tissue straight to the data source for max accuracy, whether AI or not.

Any insight from AI, whether as the summary or the ranked report, has links that lead to the calculations from the raw data. The calculations within the sheet themselves even have a link to raw data examples.

All the hidden intersecting metrics between two columns are calculated automatically. These calculations are used by researchers to understand how connected the data is.

An example is - 61% of "Marketing C-Suites" records were "Male". The super sheet calculates all combinations like this, even in the reverse direction like - 20% of "Male" records were "Marketing C-Suites."

Aggregates of the column values are calculated instantly as share and average findings. The aggregate data is shared as a tab in the super sheet and sent to AI for ranking.

By putting all the value aggregates in one place, you can see the full picture of your data vs what AI just wants to show you.

Any statistical errors or outliers that could impact the quality of the report is highlighted in an "Errors & Outliers Tab." There are certain levels of messy data that can obfuscate or even mislead you in a report.

An example of an error is a large amount of blanks or a primarily text heavy column with some numerical values. Outlier analysis is performed on all columns to show skewed data as well.