Security
No AI action goes out without a person's approval. Each workspace's data is kept apart, and every change is logged.
Every action an agent proposes waits in one approval queue. Nothing goes out until a person approves it, edits it or dismisses it.
Each suggestion shows what the agent noticed and where it looked, so you can check its working before you decide.
Each workspace's data is kept apart in the database by row-level security, so one company never sees another's records.
Every change is written to an audit log, with who made it and when.
Every page and API call travels over HTTPS.
Multi-factor sign-in is planned. Today you sign in with your email and password, or with Google.
Sign in with Google. Enterprise single sign-on through your organisation's own identity provider, and Microsoft sign-in, are planned.
Admins, managers, sales reps and users each see and do what their role allows.
Every change to Aurora is checked automatically for vulnerable dependencies and leaked secrets before it can ship.
Ask us anything about how Aurora handles your data, or report a security concern, and the people who build it will answer.