Cybersecurity

Microsoft Launches Perception Platform for Cyber Defense via AI Agents

Microsoft has announced the general availability of the Perception platform, which aims to automate and orchestrate cyber defense operations using AI agents specialized in threat detection, investigation, and remediation.

2026-11-17
3 min read
3 views
certi.news
Microsoft Launches Perception Platform for Cyber Defense via AI Agents

During the Ignite conference, Microsoft announced that its new platform, Perception, has moved to general availability. The platform aims to transform how security operations are managed by orchestrating AI agents to perform complex tasks, including detecting cyber threats, investigating their causes, and taking immediate action to remediate them, while ensuring that strategic and sensitive decisions remain under human oversight.

Defensive Orchestration via Three Virtual Teams

The Perception platform relies on an operational model that mimics the structure of traditional cybersecurity teams, coordinating work between three categories of intelligent agents:

  • Red Team: Responsible for simulating cyberattacks and proactively searching for system vulnerabilities.
  • Blue Team: Focuses on examining security alerts and investigating detected threats.
  • Green Team: Works on remediating security issues and strengthening protection measures to prevent the recurrence of vulnerabilities.

These teams operate through interconnected workflows that exchange security information and data derived from the organization's environment, allowing users to define specific security goals or select pre-built workflows for the agents to execute using general or specialized AI models.

What Changes Practically?

The launch of Perception represents a shift from the stage of "conversational assistance" to "autonomous orchestration." While the Security Copilot product acts as an assistant relying on a chat interface to answer queries, the Perception platform acts as an orchestrator that manages agents capable of executing sequential and partially autonomous defensive tasks.

This development means reducing the Mean Time to Respond (MTTR) to incidents by automating routine steps in investigation and triage, allowing security teams to focus on more complex threats. However, Microsoft imposes strict constraints to ensure governance, granting agents the minimum necessary permissions and requiring explicit human approval before executing any critical security actions.

Available Tools and Future Plans

The tools currently available to the public include a set of agents specialized in alert triage, attack investigation, detection rule creation, and threat intelligence analysis, in addition to agents for reconnaissance and security posture prioritization. Conversely, some advanced features remain in the preview stage, such as the web application penetration testing agent, the risk reduction agent, and Defender platform integrations.

Financially, Microsoft has adopted a consumption-based pricing model, with the option for "AI Flex," and provides tools for monitoring spending and setting budgets, giving organizations flexibility in managing the costs of adopting agentic AI technologies.

News source
AITnews Arabic
Open original source ↗
c
Author

certi.news

In the same category

You may also like

View all news