ECS 2027
31/05/2027, Congress Center Düsseldorf
Düsseldorf, Germany
Prepare and govern your content for AI
Copilot doesn’t know your organization. It can read your content.
Every answer Copilot gives is grounded in what the Microsoft Graph can see: your SharePoint sites, Teams files, OneDrive folders, emails, and whatever you’ve connected through Copilot connectors. Permissions decide what Copilot can reach. The semantic index decides what it can find. Your content decides what it says.
So when Copilot hallucinates, cites an outdated policy or surfaces the salary spreadsheet nobody should see, the problem usually isn’t the AI. It’s ten years of duplicates, ownerless sites, “Everyone except external users” sharing and missing metadata.
In this session we look at Copilot through the eyes of the Graph. You’ll see how grounding actually works, why oversharing becomes oversurfacing, and why undersharing is just as harmful: content nobody can find can’t be used to give a good answer. We’ll go through the M365 toolset available today, including SharePoint Advanced Management, Restricted Content Discovery, Microsoft Purview, Microsoft 365 Archive and authoritative sources. Then we’ll cover the part no tool can do for you: the ownership, the roles and the habits that keep content trustworthy after the rollout project ends.