{"id":1673,"date":"2026-09-29T10:53:31","date_gmt":"2026-09-29T09:53:31","guid":{"rendered":"https:\/\/www.senetic.co.uk\/blog\/?p=1673"},"modified":"2026-11-12T12:16:14","modified_gmt":"2026-11-12T12:16:14","slug":"microsoft-project-perception","status":"publish","type":"post","link":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/","title":{"rendered":"Security in the age of AI agents: Microsoft Project Perception and the new cyber stack"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\"><strong>Artificial intelligence is changing not only how businesses operate, but also the nature of cyber threats.<\/strong> AI systems can analyse information, make decisions and perform actions without direct human involvement. This gives cybercriminals new opportunities to automate attacks, while organisations need security measures that can also operate faster and with greater context.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Microsoft is advancing a new AI agent-based cybersecurity concept in this direction. One of its most important components is <strong>Project Perception<\/strong> \u2013 a system designed to continuously detect risk, analyse threats and take protective action.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">What does this mean for businesses? And why is AI security starting to require a fundamentally different approach from traditional infrastructure protection?<\/p>\n\n\n\n<!--more-->\n\n\n\n<h2><b><span data-contrast=\"auto\">Key points at a glance<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h2>\n<ul>\n<li><span data-contrast=\"auto\">Project Perception is a concept for an agent-based security system designed to continuously detect risk, analyse it and take protective action.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">The system uses three groups of specialised agents: Red Team, Blue Team and Green Team.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Microsoft is developing a so-called cyber stack that combines data from across the IT environment, security context, AI models, agents and action execution mechanisms.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">One emerging threat is prompt injection, an attempt to manipulate an AI system using specially crafted instructions.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Microsoft is also developing AI protection within Defender, Microsoft Agent 365, Entra and Purview.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Project Perception entered public preview on 3 August 2026, according to Microsoft. Microsoft \u2013 Rethinking security for the age of AI<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<h2><b><span data-contrast=\"auto\">Why does AI security require a new approach?<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\"><a href=\"https:\/\/www.senetic.co.uk\/blog\/?p=1633\">AI is changing not only how users work<\/a>. It is also changing the scale and pace of potential attacks.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">A traditional cyberattack often requires time for the attacker to analyse the infrastructure, prepare a campaign, identify vulnerabilities and adapt subsequent stages of the attack. AI models can automate some of these tasks and enable large volumes of information to be analysed faster.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Microsoft notes that as autonomous systems evolve, the number of elements an organisation must secure also increases. This now applies not only to computers, servers and user accounts, but also to AI applications, agents, the data used by models and the relationships between these elements. Microsoft \u2013 Rethinking security for the age of AI<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"auto\">What is Microsoft Project Perception?<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">Project Perception is an agent-based security system being developed by Microsoft for environments where AI is part of an organisation&#8217;s infrastructure.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Its purpose is to bring together several elements:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ul>\n<li><span data-contrast=\"auto\">security signals,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">information about the organisation&#8217;s environment,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">AI models,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">specialised agents,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">mechanisms for executing protective actions.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">In practice, this means moving away from individual tools operating independently towards a system that can combine information from multiple areas to create a unified view of the security situation.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Microsoft announced the public preview of Project Perception from 3 August 2026. Microsoft \u2013 Rethinking security for the age of AI<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"auto\">Red Team, Blue Team and Green Team \u2013 how do Project Perception agents work?<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">One of the most interesting aspects of Project Perception is the division of agents into three groups.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">Red Team \u2013 identifying weaknesses<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Red Team agents are intended to operate like a team conducting controlled attacks.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Their role is to identify potential paths to compromising the environment before a real attacker exploits them.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">This may involve analysing:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ul>\n<li><span data-contrast=\"auto\">software vulnerabilities,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">misconfigurations,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">excessive permissions,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">dependencies between systems,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">potential paths an attacker could use to move through the infrastructure.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">The key is to consider the chain of dependencies rather than a single alert in isolation.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">For example, a vulnerability in an application may be of limited significance on its own. However, if the application has access to sensitive data, uses a privileged account and is connected to other systems, it becomes far more significant to the security of the entire organisation.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<h3><b><span data-contrast=\"auto\">Blue Team \u2013 analysis and assessment of actual risk<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h3>\n<p><span data-contrast=\"auto\">Blue Team agents are intended to analyse signals relating to potential threats and combine them with the available context.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">This is particularly important in large organisations, where security systems can generate enormous numbers of events.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">An alert alone often does not indicate whether the organisation is actually at risk.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Relevant factors may include:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ul>\n<li><span data-contrast=\"auto\">user identity,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">the device from which the connection was made,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">the location and time of the event,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">the application in use,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">previous account activity,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">the user&#8217;s access to data,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">relationships between systems,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">previous incidents,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">information about current cybercriminal campaigns.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">An agent can therefore analyse an event in a broader context rather than treating each signal as an isolated case.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<h3><b><span data-contrast=\"auto\">Green Team \u2013 moving from analysis to action<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h3>\n<p><span data-contrast=\"auto\">The biggest difference emerges at the next stage.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Green Team agents are intended to carry out remediation actions and strengthen security controls.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Depending on the scenario, this may include:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ul>\n<li><span data-contrast=\"auto\">restricting access,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">changing the configuration,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">strengthening the security policy,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">responding to an identified vulnerability,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">securing a specific asset.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">In this way, the three groups of agents form a closed security loop:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Red Team \u2192 identifies a weakness \u2192 Blue Team \u2192 assesses the risk \u2192 Green Team \u2192 takes action \u2192 the system analyses the environment again.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Microsoft describes this model as a continuous process of discovering, assessing and improving security. Microsoft \u2013 Rethinking security for the age of AI<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"auto\">The new cyber stack: what does AI-based security consist of?<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">Project Perception is part of a broader concept described by Microsoft as a <\/span><b><span data-contrast=\"auto\">cyber stack<\/span><\/b><span data-contrast=\"auto\">.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">It can be simplified into several layers.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ol>\n<li>\n<h3><b><span data-contrast=\"auto\"> Signals \u2013 data and signals<\/span><\/b><\/h3>\n<\/li>\n<\/ol>\n<p><span data-contrast=\"auto\">The first layer provides information about what is happening in the environment.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Sources may include:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ul>\n<li><span data-contrast=\"auto\">devices,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">identities,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">applications,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">the cloud,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">data,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">AI systems,<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">security solutions.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">The more systems an organisation operates, the more important it becomes to connect them.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ol start=\"2\">\n<li>\n<h3><b><span data-contrast=\"auto\"> Security context \u2013 security context<\/span><\/b><\/h3>\n<\/li>\n<\/ol>\n<p><span data-contrast=\"auto\">Data alone is not enough.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">The system needs to know <\/span><b><span data-contrast=\"auto\">what a given event means for a specific organisation<\/span><\/b><span data-contrast=\"auto\">.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">For example, a sign-in from an unusual location may be normal behaviour for an employee travelling on business. The same event involving a privileged account may require an entirely different response.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Context therefore makes it possible to correlate information about assets, users, relationships, activity and risk.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Microsoft describes security context as a continuously updated view of the environment that enables agents to analyse situations more effectively. <\/span><a href=\"https:\/\/blogs.microsoft.com\/blog\/2026\/07\/27\/rethinking-security-for-the-age-of-ai\/?utm_source=chatgpt.com\"><span data-contrast=\"none\">Microsoft \u2013 Rethinking security for the age of AI<\/span><\/a><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ol start=\"3\">\n<li>\n<h3><b><span data-contrast=\"auto\"> Models \u2013 AI models<\/span><\/b><\/h3>\n<\/li>\n<\/ol>\n<p><span data-contrast=\"auto\">The next component is models responsible for reasoning and analysis.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Microsoft highlights a <\/span><b><span data-contrast=\"auto\">multi-model<\/span><\/b><span data-contrast=\"auto\"> approach, where not every problem needs to be solved by a single, largest model.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">A faster, less expensive model can be used for simpler tasks, while more complex analysis can be assigned to a model with stronger reasoning capabilities.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">This approach can affect both performance and the cost of the overall system.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ol start=\"4\">\n<li>\n<h3><b><span data-contrast=\"auto\"> Agents \u2013 specialised agents<\/span><\/b><\/h3>\n<\/li>\n<\/ol>\n<p><span data-contrast=\"auto\">Agents perform specific tasks.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Rather than relying on a single \u201csuper-agent\u201d responsible for everything, a system can use multiple agents specialised in specific functions.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ol start=\"5\">\n<li>\n<h3><b><span data-contrast=\"auto\"> Actuators \u2013 taking action<\/span><\/b><\/h3>\n<\/li>\n<\/ol>\n<p><span data-contrast=\"auto\">The final layer connects analysis with real-world response.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">This is where a system decision can be translated into a specific protective action.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Microsoft emphasises that this layer is deeply integrated with Microsoft Security solutions. <\/span><a href=\"https:\/\/blogs.microsoft.com\/blog\/2026\/07\/27\/rethinking-security-for-the-age-of-ai\/?utm_source=chatgpt.com\"><span data-contrast=\"none\">Microsoft \u2013 Rethinking security for the age of AI<\/span><\/a><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"auto\">Prompt injection \u2013 a new threat to AI systems<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">The growth of AI agents also creates new classes of threats.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">One of them is <\/span><b><span data-contrast=\"auto\">prompt injection<\/span><\/b><span data-contrast=\"auto\">.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Put simply, it is an attempt to influence a model\u2019s behaviour using carefully crafted instructions.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">The issue becomes particularly significant when AI not only answers questions but also has access to data and tools and can take action.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Imagine an agent that analyses email.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">An ordinary message may contain a hidden or explicit instruction directed at the model. If the agent treats it as a command rather than untrusted content, its behaviour could potentially be manipulated.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">This is why AI security cannot be limited to protecting the model itself.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">The following must also be secured:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">model + data + context + tools + permissions + how actions are performed.<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">In July 2026, Microsoft announced that it was developing prompt injection protection in Microsoft Defender to identify and isolate email messages containing malicious instructions for AI systems before they are delivered to users. The feature was presented as a preview solution. <\/span><a href=\"https:\/\/www.microsoft.com\/en-us\/security\/blog\/2026\/07\/30\/whats-new-in-microsoft-security-july-2026\/?utm_source=chatgpt.com\"><span data-contrast=\"none\">Microsoft Security \u2013 What\u2019s new in Microsoft Security: July 2026<\/span><\/a><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"auto\">An AI agent needs its own security layer<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">Introducing an AI agent into a company should not end with the question:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">\u201cDoes the agent work?\u201d<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">The following questions are equally important:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ul>\n<li><span data-contrast=\"auto\">What data can it access?<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Which applications can it use?<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">What actions can it perform independently?<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">What permissions does it have?<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Is its activity monitored?<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">What happens when suspicious activity is detected?<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<li><span data-contrast=\"auto\">Can its access be revoked quickly?<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/li>\n<\/ul>\n<p><span data-contrast=\"auto\">This is why Microsoft is developing <\/span><b><span data-contrast=\"auto\">Microsoft Agent 365<\/span><\/b><span data-contrast=\"auto\">.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">According to Microsoft, the new Defender capabilities are intended to combine security posture assessment and runtime protection for agents operating, among other environments, in Microsoft Foundry, Microsoft Copilot Studio and third-party agent management environments. <\/span><a href=\"https:\/\/www.microsoft.com\/en-us\/security\/blog\/2026\/07\/30\/whats-new-in-microsoft-security-july-2026\/?utm_source=chatgpt.com\"><span data-contrast=\"none\">Microsoft Security \u2013 What\u2019s new in Microsoft Security: July 2026<\/span><\/a><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">In practice, this shifts the focus from user and device security towards <\/span><b><span data-contrast=\"auto\">the security of digital workers<\/span><\/b><span data-contrast=\"auto\">, meaning agents that perform tasks within the organisation.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"auto\">Microsoft Defender: protecting the entire AI environment<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">The development of Project Perception does not mean that traditional security solutions are becoming less important.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Quite the opposite.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">AI must be incorporated into the existing cybersecurity architecture.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Microsoft is expanding the capabilities of <a href=\"https:\/\/www.senetic.co.uk\/blog\/?p=1613\">Defender<\/a>, integrating information from different areas of the environment and using AI in SecOps processes. In its July update, Microsoft highlighted, among other developments, the evolution of Threat Intelligence Agent and the integration of additional analytics and automation features. <\/span><a href=\"https:\/\/www.microsoft.com\/en-us\/security\/blog\/2026\/07\/30\/whats-new-in-microsoft-security-july-2026\/?utm_source=chatgpt.com\"><span data-contrast=\"none\">Microsoft Security \u2013 What\u2019s new in Microsoft Security: July 2026<\/span><\/a><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">This is important because AI security is not a separate project detached from the rest of the infrastructure.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">An AI agent uses:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">identities \u2192 devices \u2192 applications \u2192 data \u2192 networks \u2192 the cloud.<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">If any of these elements is inadequately secured, it can become part of a potential attack chain.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"auto\">Can AI protect against AI?<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">This is one of the most important questions surrounding the development of Project Perception.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Microsoft\u2019s answer is broadly affirmative: since AI enhances attackers\u2019 capabilities, similar technologies should be used by security teams.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">This does not mean, however, that people are no longer needed.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Microsoft emphasises the concept of <\/span><b><span data-contrast=\"auto\">human in control<\/span><\/b><span data-contrast=\"auto\"> \u2013 AI systems should automate analysis and actions, but people should retain control over key security processes. <\/span><a href=\"https:\/\/blogs.microsoft.com\/blog\/2026\/07\/27\/rethinking-security-for-the-age-of-ai\/?utm_source=chatgpt.com\"><span data-contrast=\"none\">Microsoft \u2013 Rethinking security for the age of AI<\/span><\/a><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">This is particularly important for high-impact actions.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Automatically blocking a suspicious process may be relatively safe. Automatically revoking an entire department\u2019s access to a critical business system requires much greater oversight.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">This is why the future of cybersecurity will probably not mean:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">\u201cAI instead of people\u201d.<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Rather:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">\u201cAI does more of the work; people make decisions where judgement and accountability are required\u201d.<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">How can your organisation prepare for AI security?<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\"><a href=\"https:\/\/www.senetic.co.uk\/blog\/?p=1661\">AI implementation<\/a> should begin with more than just selecting a model or application.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Preparing the right security model is equally important.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ol>\n<li><b><span data-contrast=\"auto\"> Identify all AI use cases<\/span><\/b><\/li>\n<\/ol>\n<p><span data-contrast=\"auto\">Do not limit yourself to officially deployed tools.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Also review which AI services employees use independently.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ol start=\"2\">\n<li><b><span data-contrast=\"auto\"> Identify agents and their permissions<\/span><\/b><\/li>\n<\/ol>\n<p><span data-contrast=\"auto\">Each agent should have a clearly defined scope of access.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">The principle should be simple:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">an agent receives only the permissions it genuinely needs.<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ol start=\"3\">\n<li><b><span data-contrast=\"auto\"> Secure your data<\/span><\/b><\/li>\n<\/ol>\n<p><span data-contrast=\"auto\">Define which data may be sent to AI models and external applications.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ol start=\"4\">\n<li><b><span data-contrast=\"auto\"> Monitor activity<\/span><\/b><\/li>\n<\/ol>\n<p><span data-contrast=\"auto\">Knowing that an agent exists is not enough.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">You must also know which data it uses, which applications it communicates with and what actions it performs.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ol start=\"5\">\n<li><b><span data-contrast=\"auto\"> Prepare defences against prompt injection<\/span><\/b><\/li>\n<\/ol>\n<p><span data-contrast=\"auto\">Instructions from untrusted sources should be treated as potentially dangerous data, rather than automatically as commands for the model.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ol start=\"6\">\n<li><b><span data-contrast=\"auto\"> Plan incident response<\/span><\/b><\/li>\n<\/ol>\n<p><span data-contrast=\"auto\">If an account, agent or application is compromised, the organisation should know how to quickly restrict access and stop the potential spread of the threat.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"auto\">7 questions to ask before deploying AI agents<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">Before deploying autonomous agents in your organisation, it is worth answering several fundamental questions:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<ol>\n<li><b><span data-contrast=\"auto\"> What data will the agent have access to?<\/span><\/b><\/li>\n<li><b><span data-contrast=\"auto\"> What actions will it be able to perform without human involvement?<\/span><\/b><\/li>\n<li><b><span data-contrast=\"auto\"> What permissions will it receive?<\/span><\/b><\/li>\n<li><b><span data-contrast=\"auto\"> Will every agent activity be logged?<\/span><\/b><\/li>\n<li><b><span data-contrast=\"auto\"> How will we detect prompt injection or other manipulation attempts?<\/span><\/b><\/li>\n<li><b><span data-contrast=\"auto\"> What will we do if an agent is taken over?<\/span><\/b><\/li>\n<li><b><span data-contrast=\"auto\"> Can we quickly revoke the agent&#8217;s access to data and applications?<\/span><\/b><\/li>\n<\/ol>\n<p><span data-contrast=\"auto\">If the organisation does not have a clear answer to any of these questions, it should first strengthen its security foundations before increasing AI autonomy.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<h2><b><span data-contrast=\"auto\">Summary<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/h2>\n<p><span data-contrast=\"auto\">The AI era is also changing how organisations need to think about cybersecurity.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Organisations must protect not only users, devices, applications and data, but also <\/span><b><span data-contrast=\"auto\">AI models and autonomous agents<\/span><\/b><span data-contrast=\"auto\">, which can perform tasks and take actions within the organisation&#8217;s environment.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">Project Perception is Microsoft&#8217;s response to this shift. Rather than a single tool, it offers a concept of a <\/span><b><span data-contrast=\"auto\">continuous, agent-based security system<\/span><\/b><span data-contrast=\"auto\">, combining data, context, AI models, specialised agents and automated response mechanisms.<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">At the same time, AI security requires a broader perspective than the model alone. The key elements are:<\/span><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><b><span data-contrast=\"auto\">identity + permissions + data + applications + agents + monitoring + automated response.<\/span><\/b><span data-ccp-props=\"{}\">\u00a0<\/span><\/p>\n<p><span data-contrast=\"auto\">How these elements are combined will determine whether AI becomes a source of new threats for the organisation or one of its most important tools for reducing them.<br \/><br \/><\/span><\/p>\n<p><img decoding=\"async\" class=\"emoji\" role=\"img\" draggable=\"false\" src=\"https:\/\/s.w.org\/images\/core\/emoji\/17.0.2\/svg\/1f680.svg\" alt=\"\ud83d\ude80\" \/>\u00a0<strong>Would you like to learn more about Microsoft cybersecurity tools? <\/strong>Our team of experts will be happy to answer your questions and help you choose the best solution:\u00a0<\/p>\n<p><strong><img decoding=\"async\" class=\"emoji\" role=\"img\" draggable=\"false\" src=\"https:\/\/s.w.org\/images\/core\/emoji\/17.0.2\/svg\/1f4de.svg\" alt=\"\ud83d\udcde\" \/><\/strong><strong> +44 (0) 20 3666 5846\u00a0 \u00a0<\/strong><\/p>\n<p><strong><img decoding=\"async\" class=\"emoji\" role=\"img\" draggable=\"false\" src=\"https:\/\/s.w.org\/images\/core\/emoji\/17.0.2\/svg\/1f4e7.svg\" alt=\"\ud83d\udce7\" \/><\/strong><strong>\u00a0<\/strong><a href=\"mailto:microsoft@senetic.com\"><strong>microsoft@senetic.com<\/strong><\/a><strong>\u00a0\u00a0<\/strong><\/p>\n<p>You can also visit our website for more information about the products. Explore it and find out more:\u00a0<a href=\"https:\/\/www.senetic.co.uk\/microsoft365\/\">MICROSOFT<\/a><\/p>\n<p><span data-contrast=\"auto\">\u00a0<\/span><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Artificial intelligence is changing not only how businesses operate, but also the nature of cyber threats. AI systems can analyse information, make decisions and perform actions without direct human involvement. This gives cybercriminals new opportunities to automate attacks, while organisations need security measures that can also operate faster and with greater context.&nbsp; Microsoft is advancing<\/p>\n","protected":false},"author":11,"featured_media":1674,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_yoast_wpseo_meta-robots-noindex":"","_yoast_wpseo_meta-robots-nofollow":"","_yoast_wpseo_meta-robots-adv":"","_yoast_wpseo_bctitle":"","_yoast_wpseo_canonical":"","_yoast_wpseo_opengraph-title":"","_yoast_wpseo_opengraph-description":"","_yoast_wpseo_opengraph-image":"","_yoast_wpseo_twitter-title":"","_yoast_wpseo_twitter-description":"","_yoast_wpseo_twitter-image":"","footnotes":""},"categories":[4,54,160,67,80,161],"tags":[175,9,57,99],"class_list":["post-1673","post","type-post","status-publish","format-standard","has-post-thumbnail","category-business-software","category-cloud-microsoft","category-cloud-software","category-it-security","category-news","category-security","tag-ai","tag-microsoft","tag-microsoft-365","tag-security"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v28.8 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>Microsoft Project Perception and AI cyber security<\/title>\n<meta name=\"description\" content=\"Discover how Microsoft Project Perception uses specialised AI agents, security context and signals to detect risks and support protective action.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/\" \/>\n<meta property=\"og:locale\" content=\"en_GB\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Microsoft Project Perception and AI cyber security\" \/>\n<meta property=\"og:description\" content=\"Discover how Microsoft Project Perception uses specialised AI agents, security context and signals to detect risks and support protective action.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/\" \/>\n<meta property=\"og:site_name\" content=\"Senetic Blog IT - Microsoft, Office 365, Azure, Ubiquiti, MikroTik\" \/>\n<meta property=\"article:published_time\" content=\"2026-09-29T09:53:31+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-11-12T12:16:14+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.senetic.co.uk\/blog\/app\/uploads\/2026\/09\/Formatka-na-bloga-29-1-1.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1200\" \/>\n\t<meta property=\"og:image:height\" content=\"800\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Marcin Halas\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Marcin Halas\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"11 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/\"},\"author\":{\"name\":\"Marcin Halas\",\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/#\\\/schema\\\/person\\\/3b154031a7d63ac577831c169162d219\"},\"headline\":\"Security in the age of AI agents: Microsoft Project Perception and the new cyber stack\",\"datePublished\":\"2026-09-29T09:53:31+00:00\",\"dateModified\":\"2026-11-12T12:16:14+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/\"},\"wordCount\":2220,\"commentCount\":0,\"image\":{\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/app\\\/uploads\\\/2026\\\/09\\\/Formatka-na-bloga-29-1-1.jpg\",\"keywords\":[\"AI\",\"Microsoft\",\"Microsoft 365\",\"security\"],\"articleSection\":[\"Business Software\",\"Cloud &amp; Microsoft\",\"Cloud &amp; Software\",\"IT Security\",\"News\",\"Security\"],\"inLanguage\":\"en-GB\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/\",\"url\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/\",\"name\":\"Microsoft Project Perception and AI cyber security\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/app\\\/uploads\\\/2026\\\/09\\\/Formatka-na-bloga-29-1-1.jpg\",\"datePublished\":\"2026-09-29T09:53:31+00:00\",\"dateModified\":\"2026-11-12T12:16:14+00:00\",\"author\":{\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/#\\\/schema\\\/person\\\/3b154031a7d63ac577831c169162d219\"},\"description\":\"Discover how Microsoft Project Perception uses specialised AI agents, security context and signals to detect risks and support protective action.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/#breadcrumb\"},\"inLanguage\":\"en-GB\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-GB\",\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/#primaryimage\",\"url\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/app\\\/uploads\\\/2026\\\/09\\\/Formatka-na-bloga-29-1-1.jpg\",\"contentUrl\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/app\\\/uploads\\\/2026\\\/09\\\/Formatka-na-bloga-29-1-1.jpg\",\"width\":1200,\"height\":800},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/1673,microsoft-project-perception\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Security in the age of AI agents: Microsoft Project Perception and the new cyber stack\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/#website\",\"url\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/\",\"name\":\"Senetic Blog IT - Microsoft, Office 365, Azure, Ubiquiti, MikroTik\",\"description\":\"\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-GB\"},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/#\\\/schema\\\/person\\\/3b154031a7d63ac577831c169162d219\",\"name\":\"Marcin Halas\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-GB\",\"@id\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/app\\\/uploads\\\/2026\\\/03\\\/Marcin-Halas-96x96.jpg\",\"url\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/app\\\/uploads\\\/2026\\\/03\\\/Marcin-Halas-96x96.jpg\",\"contentUrl\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/app\\\/uploads\\\/2026\\\/03\\\/Marcin-Halas-96x96.jpg\",\"caption\":\"Marcin Halas\"},\"description\":\"Microsoft Solutions Marketing Specialist I have been working in marketing for years, where I strive to combine an analytical approach with creative thinking. I specialize in creating and optimizing advertising campaigns. I thrive where strategy meets creativity. After work, I spend my time actively training football, cycling, or at the gym.\",\"url\":\"https:\\\/\\\/www.senetic.co.uk\\\/blog\\\/author\\\/marcin-halas\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Microsoft Project Perception and AI cyber security","description":"Discover how Microsoft Project Perception uses specialised AI agents, security context and signals to detect risks and support protective action.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/","og_locale":"en_GB","og_type":"article","og_title":"Microsoft Project Perception and AI cyber security","og_description":"Discover how Microsoft Project Perception uses specialised AI agents, security context and signals to detect risks and support protective action.","og_url":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/","og_site_name":"Senetic Blog IT - Microsoft, Office 365, Azure, Ubiquiti, MikroTik","article_published_time":"2026-09-29T09:53:31+00:00","article_modified_time":"2026-11-12T12:16:14+00:00","og_image":[{"width":1200,"height":800,"url":"https:\/\/www.senetic.co.uk\/blog\/app\/uploads\/2026\/09\/Formatka-na-bloga-29-1-1.jpg","type":"image\/jpeg"}],"author":"Marcin Halas","twitter_card":"summary_large_image","twitter_misc":{"Written by":"Marcin Halas","Est. reading time":"11 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/#article","isPartOf":{"@id":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/"},"author":{"name":"Marcin Halas","@id":"https:\/\/www.senetic.co.uk\/blog\/#\/schema\/person\/3b154031a7d63ac577831c169162d219"},"headline":"Security in the age of AI agents: Microsoft Project Perception and the new cyber stack","datePublished":"2026-09-29T09:53:31+00:00","dateModified":"2026-11-12T12:16:14+00:00","mainEntityOfPage":{"@id":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/"},"wordCount":2220,"commentCount":0,"image":{"@id":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/#primaryimage"},"thumbnailUrl":"https:\/\/www.senetic.co.uk\/blog\/app\/uploads\/2026\/09\/Formatka-na-bloga-29-1-1.jpg","keywords":["AI","Microsoft","Microsoft 365","security"],"articleSection":["Business Software","Cloud &amp; Microsoft","Cloud &amp; Software","IT Security","News","Security"],"inLanguage":"en-GB","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/","url":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/","name":"Microsoft Project Perception and AI cyber security","isPartOf":{"@id":"https:\/\/www.senetic.co.uk\/blog\/#website"},"primaryImageOfPage":{"@id":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/#primaryimage"},"image":{"@id":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/#primaryimage"},"thumbnailUrl":"https:\/\/www.senetic.co.uk\/blog\/app\/uploads\/2026\/09\/Formatka-na-bloga-29-1-1.jpg","datePublished":"2026-09-29T09:53:31+00:00","dateModified":"2026-11-12T12:16:14+00:00","author":{"@id":"https:\/\/www.senetic.co.uk\/blog\/#\/schema\/person\/3b154031a7d63ac577831c169162d219"},"description":"Discover how Microsoft Project Perception uses specialised AI agents, security context and signals to detect risks and support protective action.","breadcrumb":{"@id":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/#breadcrumb"},"inLanguage":"en-GB","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/"]}]},{"@type":"ImageObject","inLanguage":"en-GB","@id":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/#primaryimage","url":"https:\/\/www.senetic.co.uk\/blog\/app\/uploads\/2026\/09\/Formatka-na-bloga-29-1-1.jpg","contentUrl":"https:\/\/www.senetic.co.uk\/blog\/app\/uploads\/2026\/09\/Formatka-na-bloga-29-1-1.jpg","width":1200,"height":800},{"@type":"BreadcrumbList","@id":"https:\/\/www.senetic.co.uk\/blog\/1673,microsoft-project-perception\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/www.senetic.co.uk\/blog\/"},{"@type":"ListItem","position":2,"name":"Security in the age of AI agents: Microsoft Project Perception and the new cyber stack"}]},{"@type":"WebSite","@id":"https:\/\/www.senetic.co.uk\/blog\/#website","url":"https:\/\/www.senetic.co.uk\/blog\/","name":"Senetic Blog IT - Microsoft, Office 365, Azure, Ubiquiti, MikroTik","description":"","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.senetic.co.uk\/blog\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-GB"},{"@type":"Person","@id":"https:\/\/www.senetic.co.uk\/blog\/#\/schema\/person\/3b154031a7d63ac577831c169162d219","name":"Marcin Halas","image":{"@type":"ImageObject","inLanguage":"en-GB","@id":"https:\/\/www.senetic.co.uk\/blog\/app\/uploads\/2026\/03\/Marcin-Halas-96x96.jpg","url":"https:\/\/www.senetic.co.uk\/blog\/app\/uploads\/2026\/03\/Marcin-Halas-96x96.jpg","contentUrl":"https:\/\/www.senetic.co.uk\/blog\/app\/uploads\/2026\/03\/Marcin-Halas-96x96.jpg","caption":"Marcin Halas"},"description":"Microsoft Solutions Marketing Specialist I have been working in marketing for years, where I strive to combine an analytical approach with creative thinking. I specialize in creating and optimizing advertising campaigns. I thrive where strategy meets creativity. After work, I spend my time actively training football, cycling, or at the gym.","url":"https:\/\/www.senetic.co.uk\/blog\/author\/marcin-halas\/"}]}},"_links":{"self":[{"href":"https:\/\/www.senetic.co.uk\/blog\/wp-json\/wp\/v2\/posts\/1673","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.senetic.co.uk\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.senetic.co.uk\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.senetic.co.uk\/blog\/wp-json\/wp\/v2\/users\/11"}],"replies":[{"embeddable":true,"href":"https:\/\/www.senetic.co.uk\/blog\/wp-json\/wp\/v2\/comments?post=1673"}],"version-history":[{"count":2,"href":"https:\/\/www.senetic.co.uk\/blog\/wp-json\/wp\/v2\/posts\/1673\/revisions"}],"predecessor-version":[{"id":1756,"href":"https:\/\/www.senetic.co.uk\/blog\/wp-json\/wp\/v2\/posts\/1673\/revisions\/1756"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.senetic.co.uk\/blog\/wp-json\/wp\/v2\/media\/1674"}],"wp:attachment":[{"href":"https:\/\/www.senetic.co.uk\/blog\/wp-json\/wp\/v2\/media?parent=1673"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.senetic.co.uk\/blog\/wp-json\/wp\/v2\/categories?post=1673"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.senetic.co.uk\/blog\/wp-json\/wp\/v2\/tags?post=1673"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}