This is the Trace Id: 77f4e4e5897e384e070b1b69cbef04b0
Преминаване към основното съдържание Microsoft Defender Microsoft Entra Microsoft Intune Project Perception Microsoft Purview Microsoft Sentinel SIEM Вижте всички продукти Киберсигурност с подкрепа на ИИ Защита в облака Защита и управление на данните Самоличност и мрежов достъп Защита за ИИ Малки и средни фирми Интегрирани операции по сигурността Zero Trust Цени Услуги Партньори Защо Microsoft Security Осведомяване относно киберсигурността Разкази на клиенти Защита 101 Пробни версии на продукти Признание в отрасъла Microsoft Security Insider Отчет за цифровата защита на Microsoft Център за реагиране за защита Блог за Microsoft Security Събития, свързани със защитата, на Microsoft Техническа общност на Microsoft Документация Библиотека за техническо съдържание Обучение и сертификации Програма за съответствие за Microsoft Cloud Център за сигурност на Microsoft Service Trust Portal Microsoft Инициатива за защитено бъдеще Център за бизнес решения Свържете се с отдела за продажби Започнете безплатно изпробване Microsoft Security Azure Dynamics 365 Microsoft 365 Microsoft Teams Windows 365 ИИ на Microsoft Azure Space Смесена реалност Microsoft HoloLens Microsoft Viva Квантови изчисления Образование Автомобилен Финансови услуги Държавни организации Здравеопазване Производство Търговия на дребно Намерете партньор Станете партньор Мрежа от партньори Microsoft Marketplace Софтуерни фирми Блог Реклами на Microsoft Център за разработчици Документация Събития Лицензиране Microsoft Learn Microsoft Research Преглед на картата на сайта

AI-powered protection that acts at machine speed.

Read the blog
AGENTIC SYSTEM

Microsoft Perception

Specialized red, blue, and green agents coordinate in playbooks to expose gaps, investigate threats, and remediate continuously. You set the guardrails and approve high-impact actions.
Overview

Multi-agent defense

Coordinated agents. Shared intelligence. Protection that never stops.
  • Attackers now move at machine speed, and defense has to match it. Microsoft Perception orchestrates red, blue, and green team agents as one team: red agents simulate attacks, blue agents detect and triage, and green agents remediate and harden. Agents share intelligence across tools and workflows, so a finding moves straight to a fix, with fewer handoffs and less repeated analysis.
    A person sitting at a desk with a laptop.
  • Security value is measured in outcomes, not completed tasks. Perception runs comprehensive playbooks that orchestrate teams of agents towards a security outcome, grounded in agent-ready context, informed by Microsoft Threat Intelligence and broad signals across Microsoft Security.
    A person using a laptop.
  • Autonomous defense scales only when you trust it enough to turn it on. Microsoft Perception inherits your existing governance and compliance policies, runs agents with least-privilege access, and asks for approval before critical actions. Every agent action is logged and auditable, so your team delegates with confidence and stays in control.
    A close-up of hands using a phone.
WHAT’S INCLUDED

Inside Microsoft Perception

The building blocks of agentic defense — intelligence, agents, and models to power your security stack while keeping humans in charge.
Playbooks

Repeatable security outcomes

Agentic playbooks orchestrate agents toward security outcomes. Describe an objective or select a playbook, and Perception coordinates the right agents for each task.
Agents

Red, blue, & green agents

Specialized agents that expose weaknesses, detect and respond, and remediate and harden across the entire attack lifecycle.
Harness

Orchestration and governance for agents

The framework, tools, and controls that coordinate and govern every agent and model: one control plane, least-privilege access by default, approval before critical actions, and every action logged and auditable.
Models

Purpose-built cyber models

Task-optimized model selection balances quality, speed, and cost across frontier and specialized models, including the MAI-Cyber-Flash lineage, Microsoft's specialized models, purpose-built for cybersecurity tasks.
Context

Agent-ready security context

Turn enterprise-wide signals and threat intelligence into a connected, agent-ready understanding of what is happening, what is at risk, and what matters most in your environment.
Signals & actuators

From signal to action across your estate

Sensors built into the Microsoft Security portfolio feed agents real-time signals, and actuators let agents act on what they find, such as containing a threat or hardening defenses.
THE AGENTS

Meet your team of red, blue, & green agents

Three specialized agents that cover you across the full attack lifecycle, and hand off to each other automatically.
Three cybersecurity agents that simulate attacks, detect threats, and remediate vulnerabilities.
Customer stories

What customers are saying

Pricing overview

Confidently estimate and track Perception usage and spend
Pay your way
Pay as you go, or commit up front with an AI Flex Plan or MACC.
Estimate spend easily
Tools to help you understand, estimate, and plan spend, mapped to outcomes.
Stay in control
Set a budget, monitor, and manage spend with integrated cost management.
FAQ

Frequently asked questions

  • Microsoft Perception is an agentic system that brings a team of specialized AI agents to reason over your security data, tools, and workflows – delivering continuous, proactive defense with a human in control of every critical decision.

  • Microsoft Perception includes red, blue, and green agents; playbooks that coordinate agents toward a security outcome; purpose-built models, including the MAI-Cyber family; and an orchestration harness. These build on agent-ready security context and on signals, sensors, and actuators across Microsoft Security.

  • Microsoft Perception is an agentic system where specialized agents coordinate and act across your estate: AI that acts. Microsoft Security Copilot is a generative AI-assisted chat interface: AI that assists. They're complementary and priced separately.

  • An agentic security system uses AI agents to perform tasks autonomously or work together through coordinated workflows, using shared context to achieve security outcomes.

  • Red agents identify exploitable weaknesses, blue agents investigate threats, and green agents prioritize security risks and recommend actions to strengthen defenses.

  • Agentic security helps you investigate threats faster by automating repetitive tasks, correlating security signals, and coordinating response workflows.

  • You stay in control through role-based permissions, policy guardrails, human approval for high-impact actions, and auditable agent activity.

  • Start with focused, low-risk security use cases. Set clear policies and approval requirements, monitor agent activity, and expand automation as confidence grows.

  • Microsoft Perception uses consumption-based pricing, with two ways to purchase: Pay-as-you-go or via the AI Flex plan. If you have an active Microsoft Azure Consumption Commitment (MACC), your pay-as-you go and AI Flex Plan usage counts toward it.

  • Perception comes with a cost estimator tool that customers can access at set up or at any time while using the product. You can access the usage dashboard in the Defender portal under Setup & configuration > Microsoft Perception. This shows spend and outcomes delivered across agents, playbooks and users. You can also set a budget and track spend against it. 

  • AI Flex Plan is a new, simpler way to buy services across the Microsoft AI portfolio. It’s one commitment you can use across eligible Microsoft AI services, including Microsoft Foundry, Microsoft Copilot, GitHub Copilot, Microsoft Perception, Microsoft Security Copilot, and Microsoft IQ. You can shift your usage between AI services as your priorities change.

    Choose a 1- or 3-year term and pay monthly or up front.
A person using a laptop
Get started

Protect everything

Make your future more secure. Explore your security options today.

Следвайте Microsoft Security

Български (България) Поверителност на здравето на потребителите Връзка с Microsoft Поверителност Управление на бисквитките Условия за използване Търговски марки За нашите реклами EU Compliance DoCs