If the filenames are all fixed, I'd do this:
1. Create a MAP file and link all CSVs (don't need to import them).
2. Create a model table which will receive all data in the MAP file (if the fields in all files are the same, just copy and paste a table from one of the linked data sources into the MAP file, then add MFD_ID / MFD_ID_X if you want).
3. Create a query that would first do DELETE FROM <model table> and then INSERT INTO <model table> (<field1>, <field2>, ...) SELECT <field1>, <field2> ... FROM <CSV data source>::<CSV table> from each data source.
Then the process of compiling a report is: update the CSV files, open the MAP file which has them all linked, run the query, export the table to the format you want.
If the filenames are changing, there's no way to avoid manual input without writing a script, like Tim says. I'd just write a script to copy data, but you can use a combination of a query and a script to make the script part as little as possible. It's fairly straightforward either way, tell us if you need help.