Analyze a captured web table with AI

AI can help explore a captured table, but it needs the header meanings, units, and range to be clear. Treat its findings as analysis to verify, not as replacement values for the source.

Ask useful table questions

1. Capture the table with Shift + F and open it in Spreadsheet first. Remove repeated headers, totals that do not belong, and unrelated rows.
2. Check column names, units, date range, missing-value conventions, and whether the table is paginated.
3. Press A to send the checked table to the AI sidebar. Ask one task at a time: find outliers, compare periods, explain a calculation, or group categories.
4. Tell AI which columns to use and how to handle blanks, currencies, percentages, and duplicate IDs.
5. Recalculate important totals and percentages in Spreadsheet or against the source before sharing the conclusion.

What makes an analysis auditable?

Name the input range, filters, assumptions, and output columns. Ask AI to show the rows behind an observation so another person can reproduce the check.

Common questions

Why did AI invent a trend?

The capture may omit earlier rows, mix units, or contain text that looks numeric. Inspect the range and ask for evidence rows rather than accepting a narrative.

Can AI fix the table?

It can suggest transformations, but apply them in Spreadsheet and keep the original capture. Never overwrite source values without recording the rule.

What about sensitive tables?

Remove private identifiers and confirm the selected provider’s data policy before sending the table.