Your people may already be using Copilot, ChatGPT and Claude, and some of that use may involve customer or business data. Secure AI shows you which AI tools are in use and what sensitive information they can reach, then puts the right controls in place in Microsoft 365.
Staff sign up to AI tools and connect them to Microsoft 365 without IT knowing.
Copilot and agents can surface any file a person can open, including files shared widely years ago.
Customer details or commercial information can be pasted into personal AI accounts.
Without logging and retention, you can’t show what AI accessed or shared.
Our approach is aligned with Microsoft’s guidance on Security for AI and follows four steps. The steps are the same for every organisation. The tools depend on your Microsoft 365 licence, and where it doesn’t cover a capability, we may use AvePoint to fill the gap.
We check the foundations Microsoft recommends before AI is rolled out: sign-in and multi-factor authentication, devices, audit logging and licensing.
We identify the AI apps and agents in use and show where sensitive data may be overexposed, such as overshared sites, open links and guest access.
We work on the highest risks first, which may include oversharing, sensitivity labels and data loss prevention. Changes are tested before they’re enforced, and controls for AI apps depend on your licence.
We set up auditing and retention for AI activity, help you write an acceptable-use policy, and give you a roadmap for ongoing control.
Steps 1 and 2 can run on their own. We don’t change your users’ files or access; we only turn on the discovery tools needed to see your AI use. You get a readiness scorecard, a report on the AI in use and your data risks, and an executive readout. Then you decide what to do next.
AI readiness scorecard
AI usage and data risk report
A register of the protection policies we put in place
A one-page AI acceptable-use policy
An executive readout and a roadmap
Secure AI suits Microsoft 365 organisations that want to know which AI tools their people use and what those tools can reach.
Keep it safe every month with Zetta Managed Governance: we watch sharing, access and new AI apps, and report to you monthly.
Secure AI stops your data leaking into AI. Adopt AI gets your people using AI on real work. AI Business Process Automation automates the processes eating your team’s week.
| Secure AI · this page | Adopt AI | AI Business Process Automation | |
|---|---|---|---|
| In one line | Stop your data leaking into AI | Sets up the tools to get your people using AI on real work, and prove what it’s worth | Automate the processes eating your team’s week |
| The question it answers | Is our data safe from the AI our people use? | How do we get value from Copilot and Claude? | Where does your team’s week actually go? |
| What it changes | Microsoft 365 controls, with AvePoint where needed, and auditing of AI activity. | How people work day to day, with hours saved measured by task. | How the work runs: repeat steps automated with Power Automate, Copilot and low-code apps |
| Who it's mainly for | IT, security and leadership | Leadership sponsors, team managers and the business teams using the tools. | Operations leaders and process owners |
| Starts with | AI Risk Snapshot | AI Adoption Plan | Process discovery workshop |
Secure AI keeps your data out of the AI tools your people use. To get your people using AI well, see Adopt AI. To automate the work worth handing over, see AI Business Process Automation.
No changes to your users’ files or access. We turn on the discovery tools needed to see your AI use. Changes to controls only happen in the Protect step, with your approval, in test mode first.
We find the AI tools in use and connected to Microsoft 365, as far as they can be detected. Connected AI can bring those tools under Microsoft 365 controls.
No. E5 clients can use Microsoft’s own tools, such as Purview and Defender, but everyone, including E5 clients, can use AvePoint alongside them.
Book a call and we’ll show you how the AI Risk Snapshot works.
