The short answer
- On November 17, 2026, at Microsoft Ignite, Microsoft said Microsoft Perception, previously known as Project Perception, is generally available.
- Microsoft said six agents are generally available: the Alert Triage, Attack Investigation, Detection Authoring and Threat Intelligence agents, the Recon Agent, and the Posture Prioritization Agent.
- Microsoft said the Web App Pen Test Agent is in preview, the Risk Reduction Agent is in public preview, vulnerability shielding and Model Context Protocol (MCP) and API integrations are in preview, and secure AI agents and playbooks are in private preview.
- Microsoft’s Book of News says the Triage Agent now covers about 80% of alerts in a typical Microsoft Defender queue. A Microsoft Tech Community footnote says that figure is based on about 8 million alerts and “is not a guaranteed triage rate for an individual customer.”
- Microsoft’s product page says Perception uses consumption-based pricing, with pay-as-you-go or the AI Flex Plan, and is priced separately from Microsoft Security Copilot.
- SecurityWeek reported on November 18, 2026, that Perception has been in public preview inside Microsoft Defender since August.
Microsoft said on November 17, 2026, at Microsoft Ignite, that Microsoft Perception, previously known as Project Perception, is generally available. In a Microsoft Security blog post that day, Aleš Holeček, President and Chief Architect, Microsoft Security, said Perception orchestrates red, blue and green team agents through playbooks. Six of those agents are generally available. Others, including the Web App Pen Test Agent and vulnerability shielding, remain in preview.
#What changed?
Microsoft’s Ignite 2026 Book of News says Perception, “previously known as Project Perception, is now generally available, including the agents and playbooks announced in July.” A Microsoft Tech Community post dated November 17, 2026, says the company “introduced Project Perception in July”. SecurityWeek reported on November 18, 2026, that “the system has been in public preview inside Microsoft Defender since August.”
The Security blog says: “It orchestrates red, blue, and green team agents through playbooks, working together in a closed loop to deliver security outcomes.” Microsoft says: “Red team agents simulate attacks from an attacker's perspective before a breach occurs.” It also says: “Blue team agents detect, triage alerts, and investigate attacks, freeing analysts to focus on the critical decisions that matter most.” “Green team agents fix and remediate.” The Tech Community post says: “Teams can start with a playbook or describe what they need in chat, then follow the work and review results through an integrated session experience.”
Microsoft says: “Accelerate threat detection, investigation, and response with new playbooks and four blue team agents, now generally available: Alert Triage Agent, Attack Investigation Agent, Detection Authoring Agent, and Threat Intelligence Agent.” On red team agents, the Security blog describes playbooks that include “the Recon Agent, now generally available, and the Web App Pen Test Agent, now in preview”. On green team agents, it says: “Two new playbooks bring together green team agents, including the Posture Prioritization Agent, now generally available, and the Risk Reduction Agent, now in public preview.” It also says: “Coverage spans Microsoft, Amazon Web Services (AWS), and Google Cloud.”
Microsoft said the following remain in preview:
- The Web App Pen Test Agent, in preview
- The Risk Reduction Agent, in public preview
- Vulnerability shielding, in preview
- Connections through Model Context Protocol (MCP) and APIs to security information and event management (SIEM), security orchestration, automation, and response (SOAR), and IT service management tools, plus signals from AWS, Okta and Salesforce, in preview
- Secure AI agents and playbooks, in private preview, according to the Book of News
The Security blog labels vulnerability shielding as available in preview and says it “helps protect vulnerable assets from exploitation by applying targeted, temporary protection to buy defenders time before they can apply a permanent patch.”
Microsoft’s Book of News says: “The Triage Agent now covers about 80% of alerts in a typical Microsoft Defender queue.” A footnote on the Tech Community post says it is “Based on Microsoft’s analysis of approximately 8 million alerts generated across Microsoft Defender workloads (identity, endpoint, email and collaboration, cloud apps, and cloud security), covering the highest-prevalence SOC-actionable alerts of low-to-high severity.” The footnote also says: “This percentage represents aggregate coverage across the analyzed population and is not a guaranteed triage rate for an individual customer.”
On integrations, the Security blog says: “Teams can now invoke Perception playbooks in Defender, connect through Model Context Protocol (MCP) and APIs to SIEM, security orchestration, automation, and response (SOAR), and IT service management tools, and bring in signals from AWS, Okta, and Salesforce.” It then says: “These integrations are in preview.” The Book of News describes “Secure AI agents and playbooks” and marks them “In private preview.” Microsoft also said new AI Readiness Initiatives in Perception will be in preview in January 2027.
Microsoft says: “Your team sets the guardrails and signs off on high-impact actions, and automated triggers launch playbooks when an incident meets the criteria you set.” The Microsoft Perception product page says Perception “runs agents with least-privilege access, and asks for approval before critical actions.”
#How much does it cost?
The product page says: “Microsoft Perception uses consumption-based pricing, with two ways to purchase: Pay-as-you-go or via the AI Flex plan.” It says that if a customer has an active Microsoft Azure Consumption Commitment (MACC), “your pay-as-you go and AI Flex Plan usage counts toward it.” On Microsoft Security Copilot, the page says: “They're complementary and priced separately.”
The Security blog says: “New cost management dashboards and a cost estimator help you track AI usage, forecast spending, and connect costs to measurable security outcomes.” The Book of News says: “Least-privilege access, approval controls, audit logging, usage views and a cost calculator help organizations govern agent activity and plan spending.” The product page says customers can access the usage dashboard in the Defender portal under Setup & configuration, Microsoft Perception.
#Who can use it and when?
Microsoft said Perception is generally available as of November 17, 2026. SecurityWeek reported that it has been in public preview inside Microsoft Defender since August. Microsoft Learn, last updated on October 30, 2026, and still labeled “Applies to: Project Perception”, says: “Perception is organized into the following primary sections in the Microsoft Defender portal:” New Chat, Sessions, Agents, Playbooks and Insights.
#How does it compare?
The announcement follows earlier Microsoft coverage of agentic insider security controls in Perception, published October 20, 2026, when the product was still called Project Perception, and of Microsoft Agent 365.
Our analysis: The practical split is availability. Perception itself is generally available, as are six named agents. The Web App Pen Test Agent, the Risk Reduction Agent, vulnerability shielding and the MCP and API integrations remain in preview, and the Book of News places secure AI agents and playbooks in private preview. Microsoft’s about 80% figure is a coverage claim for a typical Microsoft Defender queue, based on about 8 million alerts, and Microsoft says it is not a guaranteed rate for any one customer.
#What is still unclear?
Open questions for buyers include per-unit prices, regional availability, which licenses are required, and when the Web App Pen Test Agent, the Risk Reduction Agent, vulnerability shielding and the MCP and API integrations will reach general availability.
Frequently asked questions
When did Microsoft Perception become generally available?
Microsoft said on November 17, 2026, at Microsoft Ignite, that Microsoft Perception, previously known as Project Perception, is generally available.
Which Perception agents are generally available?
Microsoft said the Alert Triage Agent, Attack Investigation Agent, Detection Authoring Agent and Threat Intelligence Agent are generally available, along with the Recon Agent and the Posture Prioritization Agent.
Which Perception features are still in preview?
Microsoft said the Web App Pen Test Agent is in preview, the Risk Reduction Agent is in public preview, vulnerability shielding and Model Context Protocol (MCP) and API integrations are in preview, and secure AI agents and playbooks are in private preview. Microsoft also said new AI Readiness Initiatives in Perception will be in preview in January 2027.
Does the Alert Triage Agent cover 80% of Defender alerts?
Microsoft’s Book of News says the Triage Agent now covers about 80% of alerts in a typical Microsoft Defender queue. A Microsoft Tech Community footnote says the figure is based on an analysis of about 8 million alerts and “is not a guaranteed triage rate for an individual customer.”
How is Microsoft Perception priced?
Microsoft’s product page says Perception uses consumption-based pricing, with two ways to purchase: pay-as-you-go or the AI Flex Plan, and that it is priced separately from Microsoft Security Copilot.
Where does Microsoft Perception run?
SecurityWeek reported that Perception has been in public preview inside Microsoft Defender since August. Microsoft Learn, last updated on October 30, 2026, and still labeled as applying to Project Perception, says Perception is organized into primary sections in the Microsoft Defender portal.
Sources
Every factual claim in this article traces back to one of these.
One email a week on AI agents and bots.
Launches, pricing and rollout changes and the guides worth reading. At most once a week, and you can leave with one click.

