This is the big question everyone here seems to be skipping over. It seems like they're using "database" in the colloquial sense and actually mean some sort of already public data that's just rate limited (for example https://www.nlrb.gov/advanced-search).

Then depending on the order of events, either scraping didn't work well enough and were given "unlimited" (not rate limited) access, or the accounts were actually denied so they fell back to scraping. Or perhaps these two things are just unrelated despite what the story is claiming.

Or maybe, even with access, they couldn't figure out how to query the actual database, so they resorted to scraping? Even with full "tenant" access, it could take some time to figure out where to look.