OK, Here's what I found out from my support case.
As of August 2019, there are now two forms of MFA policy:
1. User-specific MFA
Enabled through the account.activedirectory.windowsazure.com/usermanagement/multifactorverification.aspx page.
2. Azure Active Directory Conditional Access - Policies
Accessed via URL: (https://aad.portal.azure.com)
Click "Azure Active Directory"
Click Conditional Access
Then enable these policies:
- Baseline policy: Require MFA for admins (Preview)
- Baseline policy: Require MFA for Service Management (Preview)
The techs on the call are saying that if #2 is enabled, then you do not need to enable MFA at the end user level, because the policy will be enforced for the things that they care about.
The techs all agreed that the documentation on the partner site (https://docs.microsoft.com/en-us/partner-center/partner-security-requirements-faq?branch=isaiah%2Fse...) was inadequate to make this distinction and they included a documentation person on the call to make notes and take screen shots of the changes required to clarify the policy.
We also confirmed for the tech (and took Fiddler traces of) the Azure AD Connect logs when #1 User-Level MFA is enabled on the account used by ADConnect. Proving without a doubt that the user-level setting being enabled will break ADConnect and with it disabled it fixes ADConnect.
Finally, we also reviewed the fact that the Microsoft Security Score site is not paying attention to the Baseline Policy settings when calculating your security score. They plan to reach out to the Security Score team to have them update the score settings when the policy is configured.
