When implementingAI Access Security,tagsare applied toGenerative AI applicationsto classify them as sanctioned, tolerated, or unsanctioned. This allows organizations to enforcepolicy-based access control over AI tools, ensuring that onlyapproved applicationsare accessible while restricting or monitoring usage of untrusted or high-risk AI platforms. This classification helps security teamsmanage AI-related risks and complianceeffectively.

SincePhase 1 of the IPSec tunnel is establishedbutPhase 2 traffic is not being received, theTunnel logsin Strata Logging Serviceshould be reviewed.Tunnel logsprovide visibility into IPSec tunnel establishment, Phase 2 negotiation, and any errors or dropped packets related to encrypted traffic. This will help identify whetherESP (Encapsulating Security Payload) traffic is being blocked, mismatched security associations (SAs) exist, or if there are other issues with Prisma Access responding to Phase 2-encrypted packets.

In anExplicit Proxy deploymentwhereno agentcan be used on the endpoint,SAML (Security Assertion Markup Language)is the supported authentication method formobile users.SAMLallows authentication via anIdentity Provider (IdP)without requiring an agent on the endpoint, making it ideal for web-based authentication incloud and remote access environments. It enablesSingle Sign-On (SSO)and secure authentication without direct integration withLDAP or Kerberos, which typically require an agent or local network presence.

After configuringdomain-based split tunnelingforzoom.us, the expected behavior can be confirmed by checking therouting table on the client machine. If split tunneling is correctly configured, the traffic for zoom.usshould be routedoutsidethe GlobalProtect VPN tunnel, while other traffic follows the tunnel path.
Reviewing the routing table ensures thatonly the intended traffic is excluded from the tunnel, confirming that the split tunnel configuration is working as expected.

This option ensures thatSSL Decryptionchecks for mismatches between theServer Name Indication (SNI) fieldin the TLS handshake and theCommon Name (CN) or Subject Alternative Name (SAN) in the server certificate. If a malicious user tries to bypass content filtering by spoofing theSNI while using the real blocked website in the HTTP host header, this setting will detect the discrepancy andblock the session, preventing unauthorized access.