We wrote a basic AI usage policy, but we suspect some of our employees are secretly using unauthorized public tools to complete their work. How do we ensure our team actually has the Capacity to use AI safely under their GWC seat requirements without micromanaging them?

Question

Tyler Smith · Accepted Answer

If employees are using unauthorized AI tools behind your back, it is not a technology problem. It is an Accountability Chart problem. Under the EOS framework, every person in a seat must GWC that seat, meaning they Get it, Want it, and have the Capacity to do it. When team members use unsanctioned tools, they are demonstrating that they lack the capacity to understand the security risks they are introducing to your operating system.

Do not try to solve this by installing invasive monitoring software or writing a thirty-page rulebook that nobody will read. Instead, address the bottleneck directly. Employees turn to unauthorized tools because their current processes are cumbersome and keep them trapped in low-value tasks. You must provide them with a safe, sandboxed corporate environment where they can access advanced language models without risking proprietary data.

Once the safe infrastructure is in place, update the roles and responsibilities for each seat on your Accountability Chart. Make it clear that utilizing approved AI tools to streamline operations is encouraged, but using public, non-secure platforms is a direct violation of their seat expectations. Discuss this openly in your departmental Level 10 Meeting™ sessions. When you frame safe AI usage as a core component of having the capacity to do the job, your team will stop hiding their tools and start collaborating on how to deploy them securely.