In a recent YouTube video, Daniel Christian [MVP] examines a practical question administrators and content owners often ask: Where are Agent Builder upload files saved? His walkthrough focuses on the storage model behind the Agent Builder feature in Microsoft 365 Copilot Studio, showing exactly how uploaded documents become part of an agent’s knowledge base. As a result, viewers gain clarity on both the visible Agent Builder interface and the underlying SharePoint mechanics that hold the uploaded files. Consequently, the video helps bridge gaps between the no-code experience and tenant-level governance needs.
Daniel Christian demonstrates that uploaded files are stored in tenant-owned SharePoint Embedded containers, meaning the physical files reside inside the organization’s SharePoint environment rather than in an isolated external store. Moreover, the video explains that uploaded documents are embedded as text knowledge so the agent can use their content to ground responses, and the files show up under the Configure tab in the Agent Builder interface. Therefore, administrators should treat these containers like other SharePoint stores when considering retention, discovery, and compliance. In addition, Christian highlights that only specific file types are supported and that certain size limits apply.
Christian offers a clear explanation of the container naming and placement conventions, illustrating that an uploaded file appears as an entry tied to a container URL inside the tenant’s content storage area. Furthermore, he walks through the use of PowerShell commands to enumerate containers associated with declarative agents, which helps administrators locate hidden or embedded containers that do not appear in classic SharePoint libraries. As a result, teams can reconcile what they see in the Agent Builder UI with the actual containers in their tenant, which aids audits and troubleshooting. Additionally, the video points out that these containers can be managed using standard SharePoint admin tools once they are discovered.
Importantly, access to embedded files respects the SharePoint permissions model and any applied sensitivity labels, so unauthorized users cannot extract or view embedded content if they lack rights. However, this tenant-owned model also creates a tradeoff: while it centralizes control and integrates with existing compliance processes, it requires administrators to monitor more locations and to understand how embedded containers differ from typical team sites. Consequently, organizations must balance the convenience of no-code uploads with the need for tighter governance, particularly when sensitive documents are involved. Moreover, administrators should assess whether additional restrictions, such as limiting upload rights or automating scans, are required to reduce exposure.
The video covers practical limits and operational behavior, noting that agents can accept up to twenty files as embedded knowledge sources and support common document formats like .docx, .pdf, .xlsx, and .txt, each with defined size thresholds. Additionally, Christian demonstrates that uploads may appear gray during transfer and that direct file-level access is possible for administrators who locate the container, which impacts how deletion and retention work. Therefore, deleting an agent does not always equate to immediate or obvious file removal unless admins explicitly remove the underlying container contents or manage retention policies. As a result, teams should plan the lifecycle of agent knowledge carefully to avoid orphaned data and to ensure compliance.
As Daniel Christian points out, one challenge is discoverability: embedded containers can be less visible than regular SharePoint sites, so organizations should document where agent-related content resides and use scripts to track containers regularly. Furthermore, while embedding makes agents more capable quickly, it raises the need to balance usability against risk, prompting administrators to enforce upload policies and to use sensitivity labeling consistently. Consequently, a practical approach involves combining user training, periodic audits using PowerShell, and clear governance rules so that operators can enjoy the benefits of rapid knowledge embedding without compromising security. Finally, Christian suggests that integrating these steps into existing compliance workflows reduces surprises.
In summary, Daniel Christian’s video provides a concise and actionable explanation of how the Agent Builder stores uploaded files inside tenant-managed SharePoint Embedded containers, and he details the tools and permissions model administrators should use to manage them. Therefore, IT teams and content owners gain a clearer path for auditing, securing, and removing agent knowledge when necessary. Overall, the guidance helps organizations balance convenient agent creation with the governance and operational controls required in enterprise environments.
agent builder upload location, where are agent builder files saved, agent builder upload folder path, agent builder storage location, agent builder file directory, agent upload save location, find agent builder uploaded files, Power Virtual Agents agent builder uploads