Conditional access policies will not help you here, as they apply post (first factor) authentication (i.e. after username/password has been verified). Even if a CA policy blocks the login attempt, at this point the attacker knows credentials were successfully verified. Instead, you should block basic auth using the workload-specific controls, i.e. block SMTP auth for Exchange, block legacy auth via Set-SPOTenant, etc.
This doesn't mean you should not have a CA policy around this scenario, as it can still prevent bad actors from accessing your tenant. But it will not prevent brute-force attempts, or lockouts.