Usually messages are categorized as junk or phish if they fail to pass to necessary auth tests.
If an org was on a block list, the message would be rejected, not simply quarantined.
So for the customer that set you a test message,
Can you post the headers of the message with the personal information removed?
Look up their domain here:
What does it show for DMARC?
How about for SPF?
https://mxtoolbox.com/spf.aspx