How confident are you that your employees can't find sensitive company information with AI? Not somewhat confident. Not reasonably confident. Completely confident. Because if you're not sure, there's a simple test you can run in the next 15 minutes that might tell you more about your business’ AI readiness than months of meetings, strategy documents, and governance discussions.
Generative AI is changing how employees work. People use ChatGPT, Claude, Microsoft 365 Copilot and other AI tools to analyze data, summarize documents, write content and solve problems. But there's a growing data security question businesses need to answer: Are employees putting sensitive company information into AI tools?
Microsoft is addressing AI Agent Data Security with a new VS Code skill called run-assert-eval, which helps developers identify AI agent risks, measure how often they occur, apply runtime controls and test the agent again.
The AI race is often framed as a competition between models: Who has the smartest model? Who has the fastest chips? Who can build the most powerful AI?
Instead of asking, "Which Copilot license should we buy?", the better question is, "Are we actually ready for Copilot?"
What was once a straightforward choice between Microsoft 365 E3 for productivity and Microsoft 365 E5 for security has become a broader conversation about security, AI, and governance.
Microsoft has officially made Copilot Cowork generally available, and this marks an important shift in how businesses should think about AI.
Positioned as an AI collaborator embedded within the Microsoft ecosystem, Copilot Cowork introduces a more conversational approach to workflow automation—one where users describe tasks in everyday language and the AI agent translates those instructions into repeatable workflows (“skills”).
Legacy tools assume trust once connected. Modern access evaluates trust continuously.
Legacy endpoint tools were built for a slower world: quarterly release cycles, office-based networks, and a patch model that assumed you could “test for a week, then deploy to everyone.” That model struggles today—not because teams aren’t good, but because the risk is different:
New Zealand | United States | Australia
kia ora ❤ NZ | 10 Cadillac Drive, Suite 380 | Nashville, TN 37027