Blog Post

Microsoft Security Community Blog
5 MIN READ

Accelerate Zero Trust for AI: Expanded SecOps coverage and new guides to turn insights into action

TarekD's avatar
TarekD
Icon for Microsoft rankMicrosoft
Nov 17, 2026

Security leaders keep asking this question: our AI footprint is growing fast. How do we know our defenses are keeping up? Agents act on their own, at machine speed, with access to real data and real systems. That accelerates innovation, and it introduces new attack surfaces and trust boundaries for every security team. Attackers are using AI too, to move faster and at scale. 

AI introduces new attack vectors like prompt injection. But most AI-related incidents still run through gaps Zero Trust was designed to close: excessive access, exposed credentials, and unpatched weaknesses. 

That's why Zero Trust principles matter more than ever. For security teams, the Zero Trust foundation they’ve built still counts. The next step is to extend it. The hard part is knowing which paths are still open and proving they're narrowing over time. 

The Zero Trust Assessment is built for exactly that. Today we’re releasing Zero Trust Assessment latest version, which extends our recent expansion of Zero Trust for AI and reflects what we’ve heard from customers and partners since. Here is what’s new: 

  • Detect what prevention misses: Deeper Security Operations coverage, with new checks that measure readiness to detect and respond when prevention isn't enough. 
  • Measure progress over time, at enterprise scale: Performance improvements that help large, complex tenants run a complete assessment, plus certificate-based sign-in for scheduled runs that track progress over time. 
  • Turn findings into a roadmap: Direct import of assessment results into the Zero Trust Workshop, so every roadmap task starts with your tenant's own findings. 

These updates make Zero Trust progress easier to track, and the results are measurable. In a May 2026 Forrester Consulting Total Economic Impact™ study, a global NGO reported significant gains in identity maturity, confirmed by independent Zero Trust maturity assessments. Zero Trust-aligned detection and response playbooks also cut its mean time to respond, and it could now measure response metrics it couldn’t track before. 

Deeper Security Operations coverage: Assume breach applies to agents, too 

Preventive controls reduce the chance of unauthorized access. Security operations address the harder question: how do you detect and respond to suspicious activity that gets through? 

With AI agents, that question becomes more urgent. An agent may sign in with valid credentials, hold legitimate permissions, and still act in unexpected or harmful ways. Preventive controls remain essential, but they can't anticipate every action after access is granted. Security operations provide the telemetry, behavioral detection, investigation, and response capabilities needed to identify and contain misuse. AI's speed raises the stakes: suspicious activity can unfold faster, leaving defenders less time to act. That's why assume breach isn't a pessimistic principle. It's a practical one, and strong security operations put it into practice. 

Figure 1: New Security Operations assessment

The latest release adds 26 new Security Operations checks that expand on the SecOps coverage introduced in August, such as privileged account protection, ransomware prevention through attack surface reduction rules, and security log retention. They assess configurations across identity and privileged access, email and endpoint protection, cloud app session controls, and Microsoft Sentinel capabilities, including data collection, threat intelligence, hunting, retention, and automated response. Together, these checks help security teams identify gaps in prevention, detection, investigation, and response, and give the partners who support them a clear starting point for customer conversations. 

 

"Microsoft gave us the telemetry we needed to see the full picture. Our Security teams became empowered versus challenged by the complex patchwork of systems that had been in place." 

-- Weston Maggetti, Platform Manager, Ford Motor Company 

 

Built for enterprise scale and long-term progress  

The organizations with the most to protect are often the hardest to assess, and in the AI era, their environments change faster than any single assessment can capture. This release addresses both. Performance improvements make large tenants easier to assess at real enterprise scale: one organization recently completed a full assessment of a tenant with more than 500,000 member users and close to 4 million guest users. New settings also give teams control over which checks to run and how long each one can take.  

With app-only, certificate-based authentication, the assessment runs automatically without anyone signing in for supported services, on a schedule that matches your operating rhythm, from monthly reviews to every change window. Partners can use the same automated, scheduled approach to deliver recurring assessments for their customers: baseline once, then report progress on a regular cadence. Because every run is consistent, security teams can compare results over time and show leadership the progress made, the risk that remains, and what to tackle next. Assessment results stay in your environment. By default, the tool sends a basic usage event that includes your tenant ID but no assessment results, and you can turn it off with a single setting. 

Figure 2: Zero Trust Posture View
Turn findings into lasting, measurable progress 

Knowing where you stand only matters if you can act on it. With this release, you can import Zero Trust Assessment results into the Zero Trust Workshop. Your tenant's findings then appear on the matching tasks, giving your team a head start on a prioritized 12- to 24-month roadmap. Partners can import their customers’ results the same way and use the findings to anchor workshops and shape the roadmap with their customers. Repeatable assessments give you measurable inputs to set a baseline, prioritize improvements, and track progress over time. For a broader, multiyear security transformation, the Security Adoption Framework (SAF) uses the same findings to connect business outcomes, security disciplines, people, process, and technology in one coordinated plan. 

Get started 

Learn more about Microsoft Security solutions on our website and bookmark the Microsoft Security blog for expert insights on security matters. Follow us on LinkedIn (Microsoft Security) and X (@MSFTSecurity) for the latest cybersecurity news and updates. 

Updated Nov 16, 2026
Version 1.0