Use case
Make sure a file is right before anyone acts on it
One duplicated export or a mismatched total can send a team in the wrong direction. Most people find the problem too late.
For: Analysts, finance, and anyone who shares numbers with others
Why it matters
A wrong number, once shared, is hard to take back. A quick check up front costs far less than a correction later.
Ask in order
Three questions that answer it
Ask them in this order. Each answer sets up the next question.
How many rows are missing a date or amount, by column?
Finds the gaps that quietly skew totals and averages.
Are there duplicate deal or order IDs?
Catches rows counted twice by a repeated export.
Does total revenue in this export match the ledger for the same month?
Confirms the file agrees with a number you already trust.
Then follow up
Ask about the chart, not the whole file again
Follow-up questions build on what is already on the canvas, so you don't restate the dataset or the time window.
Show the duplicate rows side by side.
Lets you decide which copy to keep before you clean the file.
Data you need
The export you plan to use, plus one total you already know from another source.
Decision it drives
Whether the file is ready to share, or needs cleaning first.
Check the answer
Check every number before you act on it
Open the query behind any chart to see how a number was worked out and which figures it came from. When a question could mean two things, such as gross or net revenue, DashboardGPT asks you which one you mean.
Open the query behind each check to see exactly which rows it flagged.
FAQ
Common questions
More use cases
Run these three questions on your own export
The playground runs on demo SaaS, e-commerce and website data, so it shows how answers work rather than these exact questions. To ask them of your own export, create an account and upload it. Every new account starts on a free 15-day Team trial, with no card required.