Mailgun's validation API, presumably the underpinnings of Pangram's, returns more than a simple yes/no validity. My educated guess is that this is part of figuring out all of those extra fields.
* https://mailgun.com/products/validate/
* https://documentation.mailgun.com/docs/validate/oas/openapi-...
This is a good bet, particularly this:
> Catch email addresses that have turned into honey pots
> Make smart decisions on who you should and shouldn’t send to using our risk score
Identifying honeypots is tricky business. Sending something that looks like obvious spam from random burner domains and seeing if it still gets delivered is not a bad way to do it.