The tools you cannot see
Most AI use at work happens on personal accounts. It does not appear in your SSO logs, your DLP rules or your vendor register — which is exactly why it is unmeasured.
Free · No login · 60 seconds
How much shadow AI is your organisation carrying?
Five questions from a workforce instrument built for exactly this. You get an AI Risk Exposure band, a SafeScore, and one honest sentence about the biggest lever you have.
This is the human layer of your AI security posture — not a penetration test, and not a replacement for data-loss prevention.
Why this is not in your security stack already
The data that leaves with it
Client names, financial figures, unreleased code, personal data. Not through malice — through a deadline at 9pm and an approved tool that was not good enough.
The gap nobody reports
People do not volunteer that they pasted a contract into a free chatbot. An anonymous instrument gets an answer a compliance survey never will.
What the full instrument does
AI Safe Use asks your whole workforce 28 questions, anonymously. It reports by department and role, never by person, and never below five responses in a group — so it can tell you where to invest without ever telling you who to blame. Your people get a private map of their own habits and a few things to do differently. You get an AI Cyber Risk Register: measured exposure, the provisioning gap driving it, and a policy reality check.
AI Safe Use informs no decision about any individual’s employment. That is a design constraint of the instrument, not a policy we ask customers to adopt.