The YouTube video by Szymon Bochniak (365 atWork) explains how administrators can turn off SharePoint Agents in a Microsoft 365 environment. In clear steps, the presenter shows the main controls available today and demonstrates practical approaches for organizations that are not ready to deploy agents widely. Consequently, the video is useful for IT teams who need immediate, actionable guidance. Moreover, it frames the options in the context of current licensing and administrative capabilities.
SharePoint Agents are AI-powered assistants tied to SharePoint sites and surfaced through Microsoft’s Copilot functionality. They appear as .agent files on site libraries and can automate queries, workflows, and simple tasks that help users find information faster. As such, they can increase productivity and reduce repetitive work, but they also introduce new governance and compliance considerations for administrators. Therefore, understanding how they are created and stored is central to managing risks and benefits.
The video walks administrators through several practical methods to disable or limit agents. First, Szymon highlights the PowerShell cmdlet that opts a tenant out of the Copilot promotion, which effectively disables default agent behavior for eligible tenants; this approach is useful for scripted, repeatable control across many tenants. Second, he shows that deleting specific .agent files removes individual agents and that license management can prevent users from creating or running agents in the first place.
Each control option involves tradeoffs between immediacy, scope, and ongoing management effort. For example, opting out with PowerShell can quickly remove unwanted functionality at scale, yet it requires scripting skills and careful planning to avoid unintended impacts. Conversely, deleting .agent files gives fine-grained control but becomes labor-intensive in large environments and risks missing newly created agents.
The video also discusses the temporary licensing promotion that has broadened agent access, allowing more users to interact with agents under capped query limits. While this promotion encourages adoption, it also complicates governance because users without permanent Copilot licenses can still create or run agents for a limited time. Therefore, administrators must weigh the benefit of wider adoption against the need for policy, auditing, and user training to avoid accidental data exposure or compliance gaps.
Importantly, the presenter notes that there is no single tenant-wide toggle yet that universally disables agents with one click. As a result, large organizations must combine license controls, PowerShell opt-out, and file cleanup to achieve a consistent state. This multiplies the operational overhead, and consequently IT teams should prepare processes and monitoring to catch new agent files or changes. In addition, the lack of a unified control increases the importance of robust change management and coordination between security, compliance, and business units.
From a governance perspective, agents raise questions about data access, audit trails, and who can create intelligence-driven automations. Although agents can improve context continuity and productivity, they also widen the attack surface when not properly governed. Thus, organizations must balance productivity gains with security controls such as restrictive site settings, strict license assignment, and routine reviews of .agent files and permissions.
Szymon’s recommendations are pragmatic: start by auditing existing sites for .agent files and review Copilot license assignments, then use PowerShell for tenant-level changes when needed. Next, develop a governance policy that defines who may create agents, what data they may access, and how usage is monitored. Finally, train users about acceptable use and the limits of the promotion so that business teams understand both the value and the responsibilities involved.
The presenter notes that Microsoft plans to improve administrative controls over time, which should simplify management and reduce the need for multi-step workarounds. Meanwhile, administrators should monitor official Microsoft messages and service updates to catch new controls as they arrive. In the short term, the combination of license management, PowerShell opt-out, and targeted deletions remains the most practical approach to controlling SharePoint Agents.
Overall, the video by Szymon Bochniak offers a concise, operational guide for turning off SharePoint Agents and managing their rollout. It balances technical steps with governance guidance, and it highlights the tradeoffs administrators face between ease of use and control. Therefore, IT teams should treat agents as a capability that requires policy, monitoring, and clear ownership before broad deployment.
disable SharePoint agents Microsoft 365, turn off SharePoint agents M365, stop SharePoint agents Microsoft 365, disable SharePoint telemetry agents, disable SharePoint background agents PowerShell, manage SharePoint agents admin center, block SharePoint agents Microsoft 365, PowerShell disable SharePoint agents