Security & trust
Security pages usually tell you what a vendor wants you to believe. This one tells you what is actually built in, and what we deliberately do not claim yet.
Each organisation’s videos, drafts, approved versions, training records and billing are kept separate from every other organisation. One company can never see another’s data — that separation is built into the system, not an optional setting.
The AI only suggests: it drafts steps, warnings and checklist items from a task video. A named manager must review and approve before anything reaches workers. The AI can never approve, publish or change billing on its own.
Approval creates a locked version that can’t be edited afterwards; training, quizzes and checklists are tied to that version. You can always trace what a worker trained on back to the exact approved version and the person who approved it.
Task videos are locked to your organisation. Worker links are private, single-purpose links, served with strict security settings — not indexed by search engines and locked down against misuse.
Public demos run on made-up example data in a separate, throwaway demo setup. No real customer data is ever used in public demos, and we ask you not to share confidential data while you evaluate.
SOPCam Sentinel is our plan for continuous checking: automated, ongoing verification of the security and safety rules described above, with proof produced as a by-product of normal use. It is a plan we are building toward with design partners — not a current certification or audit result.