jberg7120 I think part of the issue is that these terms aren't well defined. Since there is only a single endpoint for all types of SMTP submission, there really is no systematic differentiation between the different types. Instead, a combination of factors is used to determine whether a submission is allowed. Some of these factors are visible/manageable by tenant admins via mail flow rules and connectors, but others are not (e.g. allow internal recipients only for unauthenticated submissions). That means that the definition of "Direct Send" doesn't really match what they've said here, which is essentially "Sending directly to the endpoint." -- not untrue, but also true of both IP-based connectors and authenticated smtp client submission.