Run operations in plain language with the Duall Master AI Assistant
Ask "who entered floor 4 after 10pm?" and get an answer — with every high-risk action gated by PIN re-authentication and a full audit trail.
Security operators spend too much time digging through logs across separate tools. The Duall Master AI Assistant turns that into a question: ask in plain language, get an answer grounded in real events.
Ask, don't dig
"Who entered floor 4 after 10pm last night?" or "Show me denied badges at the loading dock this week." The assistant reads from the same access, visitor, and video data your operators already work with — and it answers in English or Vietnamese.
Safe by design
The assistant never holds write credentials. Every action runs on behalf of the logged-in operator, with their permissions enforced downstream. High-risk actions add a second gate:
- Revoking a credential or disabling a user requires PIN re-authentication
- Every action is previewed, confirmed, then written to the immutable audit log
- Provider failures degrade cleanly — the assistant says "unavailable" instead of hanging a workflow
Where your data goes
For sites under data-residency policy, the AI provider can run inside the deployment — no payload leaves the perimeter. In production, the service refuses to start against a cloud model when residency is required.
Why it matters at scale
On a campus of nearly 8,000 people, or a hospital with 84 doors and 14 lifts, the answer to “who entered the server room after hours?” shouldn’t take an investigation. And for the defense and government sites that already run Duall Master on-premise, the assistant runs entirely inside the perimeter — sensitive queries never leave the site.
Keep reading
