Every company runs AI tools IT never approved. Today is amnesty: tell us — anonymously — which unsanctioned AI tools your company uses and what data goes into them. We'll hand you an honest risk grade, A through F, with the exact math shown.
Shadow AI is any AI tool employees use without IT approval — personal ChatGPT accounts, free Copilot tiers, side Claude or Gemini usage. It becomes a security risk the moment company data flows into tools nobody monitors.
You start at 100 points. Deductions:
| Factor | Deduction |
|---|---|
| Each unsanctioned AI tool in use | −10 |
| Customer data goes into the tools | −15 |
| Source code goes into the tools | −15 |
| Financials go into the tools | −10 |
| HR documents go into the tools | −10 |
| Company size 51–1,000 | −5 |
| Company size 1,001+ | −10 |
Grades: A = 90–100 · B = 75–89 · C = 60–74 · D = 40–59 · F = below 40. No hidden logic — the result page replays every deduction line by line.
Self-reported self-assessment, not an audit. Score saved on this device only — delete my answers.
Rogue-Agent Defense inventories every agent and AI tool touching your systems — sanctioned or shadow — and flags the risky ones before they leak something.
Get early access to Rogue-Agent DefenseBecause it is the fastest-growing unmonitored data flow in most companies. Employees paste contracts, code, and customer data into unapproved AI tools daily — each one a breach vector with no logging, no DLP, and no deletion guarantee. Amnesty works better than prohibition: declare the tools, grade the risk honestly, then approve safe paths. You cannot secure what you pretend is not there.
Yes. Your answers are scored by JavaScript in your own browser and stored only in your device's localStorage. Nothing is uploaded, sent, or tracked.
No. This is a self-assessment, not an audit. The grade comes from a transparent rule-based scoring formula shown on the page.
You start at 100 points. Points are deducted per unsanctioned tool, per sensitive data type entered, and by company size band.
The fastest wins: move sensitive work off personal AI accounts, get IT to approve one sanctioned tool per job, and set a simple policy people can actually follow.
Shadow AI is any AI tool employees use without IT approval — personal ChatGPT accounts, free Copilot tiers, side Claude or Gemini usage.