“We have CA everywhere”
In practice, CA coverage is uneven. Legacy protocols, service accounts, and emergency exclusions quietly create high-trust lanes.
Conditional Access is often treated as a one-time control. In practice, it degrades slowly; through exceptions, legacy paths, and emergency bypasses; until it no longer enforces risk where attackers actually operate.
Most CA failures are not configuration mistakes; they are governance failures.
Most organizations deploy Conditional Access correctly; once. Over time, business pressure, migrations, vendor needs, and incident workarounds erode enforcement. Attackers exploit the remaining allowed paths.
In practice, CA coverage is uneven. Legacy protocols, service accounts, and emergency exclusions quietly create high-trust lanes.
Token reuse, consent abuse, and legacy auth paths allow attackers to operate inside “compliant” environments without triggering CA challenges.
Leadership assumes CA mitigates identity risk; until an incident demonstrates otherwise.
Drift is caused by human decisions under pressure; not bad intent.
Migration, outages, and vendor access often introduce exclusions that persist indefinitely.
Protocols and apps that bypass CA remain active for compatibility; creating permanent weak points.
Emergency access becomes operational convenience, sidestepping CA entirely.
Multiple overlapping policies obscure intent and reduce enforceability.
Conditional Access requires lifecycle management, not just configuration.
Someone must own CA drift: reviewing exceptions, validating coverage, and removing legacy paths.
Measure what traffic is actually governed by CA and what is not.
Successful legacy logins and token usage are higher signal than blocked attempts.
Break-glass should be tested, logged, and tightly constrained; not assumed safe.