At Microsoft Ignite 2026, we’re announcing new innovations across Microsoft Defender and Microsoft Sentinel that bring security operations and threat protection closer together, help organizations get ahead of attacks, secure AI agents and strengthen protection across critical workloads.
- The foundation for agentic security: Integrated Security Operations Center (ISOC) in Microsoft Defender now includes the full breadth of SIEM capabilities and 90-days of built-in Defender data.
- Protection before the patch: Vulnerability shielding in Defender autonomously protects exposed systems before and during active attacks, giving security teams more time to deploy permanent patches to vulnerabilities.
- Protect AI agents wherever they run: Manage security posture for third-party agent platforms, prevent device code phishing and extend runtime protection for local AI agents on Windows.
- Protect people where they work: Extend defenses against browser-based attacks, malicious QR codes in Teams, and synthetic media threats.
Together, these innovations build on a shared foundation of signals, actuators, and coordinated response to help security teams move faster and stop attacks before they spread.
The foundation for agentic security
ISOC now includes the full breadth of SIEM capabilities and 90 days of built-in retention
The integrated SOC brings SIEM and threat protection together to provide the foundation for the agentic SOC. It gives people and agents shared context and the ability to see, understand, and act across the environment, so the security stack works as one system.
Building on the September 23 announcement, core ISOC capabilities in Microsoft Defender are now generally available, including case management, playbooks, workbooks and UEBA. Workspace creation and third-party ingestion are also generally available. In addition, we are expanding the preview to include the full breadth of SIEM capabilities for eligible Defender customers, while enabling existing Sentinel customers to opt into ISOC.
“What excites us most about ISOC is having a single pane of glass for security operations. Bringing SIEM and XDR together simplifies the analyst experience, while giving us an opportunity to optimize ingestion and automation costs.”
- Farhan Nadeem, Senior Security Officer, Government of Nunavut
New capabilities available in preview include 90 days of built-in Defender data retention, SOC optimization recommendations, Content Hub, detections as code, and automation audit logs.
Figure 1. The Defender home dashboard showcasing the new Integrated Security Operations Center (ISOC) experienceSimplify and scale security analytics with ingest-time data normalization
Security data comes from diverse sources and formats, making analysis at scale complex. To simplify this, Microsoft is introducing ingest-time normalization, which standardizes data at time of ingestion, thereby simplifying queries, improving performance, and enabling reuse across detections, investigations, threat hunting, analytics, and agentic scenarios. This helps teams spend less time managing data, and more time stopping threats. It is now available in preview.
Figure 2: Data normalization via parser management for query-ready security analytics.Protection before the patch
AI is accelerating vulnerability discovery and exploitation. In October 2026 alone, 1,413 CVEs were disclosed, marking it the largest Patch Tuesday on record. Attackers can exploit vulnerabilities within hours, while organizations often need days or weeks to patch. That gap creates a critical window of exposure.
Vulnerability shielding in Microsoft Defender, now in preview, helps close that gap by protecting vulnerable systems until patches are deployed.
On by default in Microsoft Defender, vulnerability shielding applies targeted, temporary protection to vulnerable systems in the gap between exposure and remediation, buying defenders time before they can apply a permanent patch. This AI-based capability reasons over the latest threat intelligence and each organization’s unique environment to choose the narrowest effective protection. During an active attack, it applies controls autonomously, cutting off new attack paths for adversaries and limiting the impact to your organization. Learn more.
Figure 3. Vulnerability shielding applied to a SQL server with a known public vulnerabilityProtect AI agents wherever they run
AI agents now operate across cloud platforms and endpoints, accessing data, calling tools like MCP servers and APIs, and taking action, often with users’ own permissions. Security teams need visibility into what agents can access and the ability to stop them when compromised or manipulated.
Assess posture for agents on third-party platforms
Defender is extending security posture assessment to agents built on Amazon Bedrock, Google Gemini, Google Vertex AI, and Anthropic Claude.
Now in preview, they appear in the same AI agent inventory as your Microsoft Copilot Studio and Microsoft Foundry agents, with risk assessments and security recommendations, so you can prioritize risk across platforms. Bring them into scope by connecting their platforms to Agent 365 through connected platforms.
Figure 4. Agent profile with security posture for Amazon Bedrock AgentProtect local agents at runtime
Building on local AI agent discovery, runtime protection for local agents is now generally available* on Windows. Defender inspects critical points in the agent loop and can block malicious instructions before the agent acts. Supported agents include Claude Code, GitHub Copilot CLI, OpenClaw, ChatGPT Desktop (Chat) with the full list here.
Each detection raises an alert with agent and session context, correlated into cases for your SOC; in Block mode, Defender also stops the action and notifies the user.
Protect AI agents from attacks on trusted identities
AI agent security goes beyond prompt injection. Attackers can exploit trusted identities and delegated permissions to manipulate agents without compromising them directly. New Microsoft Defender detections identify identity-based attacks, such as device-code phishing that compromises a developer’s GitHub account and enables misuse of GitHub Copilot CLI. Defender can detect suspicious activity and automatically disrupt attacks by disabling compromised Github accounts, revoking sessions, and triggering additional identity protections.
Learn more about this emerging attack pattern in our latest threat research blog.
Figure 5: Microsoft Defender detects a device code flow phishing attempt targeting a user’s GitHub account and disrupts the threat.Protect people where they work
Attackers continue to target the places where people work and communicate. New protections across the browser, endpoint, email, and Microsoft Teams meetings help security teams identify malicious activity earlier and investigate it alongside signals from across the environment.
Block browser lures and improve visibility into device health
Attackers increasingly use browser-based lures to trick users into running malicious commands on endpoints. Today, Microsoft Defender is introducing new capabilities to detect and block malicious copy-and-paste attacks called ClickFix attacks, which accounted for 12.8% of all intrusions observed by Microsoft Defender this year.
New capabilities in the Defender portal also improve visibility into endpoint health and provide greater control over preview features and settings policies. Learn more about Ignite 2026 Defender endpoint security announcements.
Strengthen protection against phishing and deepfakes in Microsoft Teams
As attackers increasingly use malicious QR codes and AI-generated impersonation, Defender is expanding protection across Microsoft Teams chats and detection for meetings.
In Teams chats, Defender detects malicious URLs hidden in QR codes, blurs harmful content, and displays warning banners. Analysts can investigate threats and remediation actions through advanced hunting, with no additional administrator setup required. Now in public preview.
Figure 6: Microsoft Defender detects a malicious URL in a QR code shared in a Microsoft Teams channel and blocks the potentially harmful message.Additionally, in Teams meetings, new synthetic audio and video detection helps identify potential deepfakes using certified third-party solutions. Detections are surfaced in Defender for investigation alongside other security signals. Availability varies by provider; contact participating providers to get started.
Security is at an inflection point. As AI transforms both how organizations operate and how attackers move, security must innovate just as quickly. These advancements across Microsoft Defender and Microsoft Sentinel meet this moment by bringing together security operations and threat protection, protecting people and AI agents wherever they work, and helping defenders stay ahead of emerging attacks. Together, they provide the foundation organizations need to innovate with confidence in the agentic era.
See what's next. Build what's possible.
Join us in San Francisco, November 17–20, or join online, for deep dives and practical labs to help you maximize your Microsoft Defender investments and to get more from the Microsoft capabilities you already use.
Featured Ignite sessions:
- KEY01: Microsoft Ignite opening keynote
- BRK324: The Alert Is Dead. The Blueprint for the Agentic SOC
- BRK333: How Threat Research Powers Attack Disruption in the AI Era
- BRK334: Modern Phishing Doesn't End at the Click
- BRK320: Agentic Security Unfiltered with Allie Mellen: Hype vs. Reality
- BRK340: A Zero Trust approach to securing AI runtime with Microsoft Defender