AI tools in use
Which AI apps and extensions are used on company devices, how often and by which teams.
A short, read-only review of AI usage and Microsoft 365 security. You get real numbers from your own company, a prioritised list of fixes and a call to walk through them.

Each one is answered with data from your own tenant and devices, not a generic survey.
Which AI apps and extensions are used on company devices, how often and by which teams.
Where AI is used with personal accounts that IT cannot see or control.
Types of sensitive data reaching AI tools, such as customer, financial or personal data, shown by category.
Secure Score, MFA and Conditional Access, device management, Defender and sharing settings.
Sensitivity labels, DLP rules and audit log settings that AI controls depend on.
What your current Microsoft plan can already do, and where an add-on would close a real gap.
20 minutes to understand your setup, licences and concerns, and agree the scope.
You grant a read-only role. Discovery tools are switched on where your licence supports them. Nothing changes for staff.
One to two weeks of real AI usage data plus a review of your Microsoft 365 security settings.
A prioritised report written for management and IT, and a call to agree what to fix first.
These are illustrative examples of the kind of findings the check produces. Your report is based only on your own data, with the fix for each finding and the licence it needs.
Prefer to check yourself first? Try the 10-question shadow AI risk assessment or the AI acceptable-use policy template.
A fixed-scope review that shows which AI tools your staff use, with which accounts, what kinds of company data reach them, and how well your Microsoft 365 security settings protect that data. You get a prioritised report and a walkthrough call.
No. The check uses read-only access. Discovery features are switched on with your approval and can be removed after the report.
No. The report shows AI apps, usage levels and categories of sensitive data, not individual conversations.
Usually one to two weeks from access being granted, depending on how much usage data we need to collect.
Business Premium, E3 and E5. Some discovery features need Defender for Endpoint or E5; where a feature is not licensed we use what is available and say what is missing.
You can fix the findings yourself, with your IT provider, or with us. Common next steps are AI data leak protection, a safe AI tool for staff and monthly monitoring.
It includes it. The AI Data Risk Check covers the Microsoft 365 security review plus AI tools, accounts and data shared with AI.