What is your goal?
I want to download an AirTable base with 7 tables and export it as one excel workbook with 7 sheets for each table and upload to DropBox.
What is the problem & what have you tried?
I have tried using Maia to help build it using Excel search workbooks, an API call, and AirTable search records. The excel search workbooks is not working properly to find the workbook.
I am basically starting from scratch and need advice
Screenshots (scenario flow, module settings, errors)
The module sending you in circles is Search Workbooks. It only finds an existing workbook in the connected Microsoft account. It does not create an Excel file from an Airtable base.
For a seven-sheet export, start with a clean .xlsx template in OneDrive. Create the seven worksheets once, and put an empty Excel Table with the correct headers on each worksheet. At the start of every run, use OneDrive > Copy a File to make a fresh working copy. Use the copied file in every Microsoft 365 Excel module. Do not write into the template itself, or rows from an earlier export can remain in the next file.
Build and verify one table first: Airtable > Search Records, then Microsoft 365 Excel > Add a Table Row, targeting the matching Excel Table in the copied workbook. Set the Airtable Limit deliberately because the documented default is 10. Compare the number of Airtable records returned with the number of rows written before adding the other six tables.
Do not put the seven exports on Router branches and expect them to join again. Put each fixed table mapping in its own subscenario, aggregate its row-write results to one completion output, and call those seven subscenarios one at a time from a parent scenario with Wait for the scenario to finish set to Yes. Pass the copied workbook ID into each call.
After the seventh call returns, use Microsoft 365 Excel > Download a Workbook, then map that file into Dropbox > Upload a File. Test the complete path with two small tables first. The acceptance check is one new workbook, two correctly named sheets, matching source and output row counts, and no rows carried over from the previous run. Then extend the same pattern to all seven tables.
