Exam NetSec-Architect Topic 1 Question 60 Discussion

An organization has a directive to adopt a Zero Trust framework focused on using identity and role-based access groups, device security and content inspection across all Security policies. To achieve this goal, an Enterprise License Agreement (ELA) was purchased, including Advanced Threat Prevention, IoT Security, and GlobalProtect.
The current security architecture uses Panorama to manage 60 NGFWs - a mix of PA-3240, PA-1410, and PA-440. Sites with PA-3240s host private application resources in the trust data center zone All sites have an untrust zone for internet access and a users zone for managed and unmanaged endpoint devices. A transit mesh zone exists to establish site-to-site connectivity through PAN-OS SD-WAN.
Privately hosted applications include web servers, SMB and NFS file servers and hosted Active Directory. The organization is in the process of adopting group mapping restrictions to these private applications, with daily additions of groups. It is also planning to build AI applications to assist the data teams with complex queries that will be hosted in the large offices containing data centers and is exploring hosting in the public cloud.
The organization uses on-premises Exchange, Dropbox, Zoom, and ChatGPT. There are a number of shadow SaaS applications that require further investigation. Users have been using Google Drive to upload confidential files within the organization by using their personal logins.
IoT devices on the network are associated on their own VLAN on the users zone. Using Device Security, all IoT devices have been categorized by asset profiles with medium or high confidence, policy sets imported into Panorama, and a default deny applied to the IoT networks.
The organization has rolled out SSL decryption and is using URL categorization for the majority of content filtering. Malicious categories, unknown and high-risk websites are blocked, with the remainder of sites set to alert.
Which deployment method should the architect suggest for enabling User-ID based rules, restricting or allowing access as close to the source as possible, while minimizing operational overhead?