AI ASSISTANT
Bring context to every investigation
A read-only operator assistant summarizes the policies, domains, risk activity and audit records an operator can already see. People make every decision.
Faster evidence gathering, human decisions
The assistant summarizes what an operator can already reach. It does not act on policies or traffic.
Read-only tools
- Policies
- Rules
- Domains
- Recent fingerprints
- Risk distribution
- Audit records
Within the signed-in operator's role and the active tenant.
Out of scope: change something
- Change configuration
- Create or delete records
- Flip settings
- Act on traffic
Write tools are not part of the current assistant.
Tool activity is recorded for review
Summaries inside the console
Read-only tools
Same permissions as the operator
Audit and usage oversight
From question to reviewed answer
Operator
Sign in as an Admin
Use an account with the Admin role and open the assistant if it is enabled for your tenant.
Console
Assistant in the console
A bounded question
Runs in your session, with your role and the active tenant.
Tools
Read what you can already see
Role-checked reads
Policies, rules, domains, recent fingerprints, risk distribution and audit records.
Console
Summary with links
What the tools found
Links to the policy, domain, risk and audit views for follow-up.
Your team
Verify and decide
Check the summary against the product records. Operational decisions stay with your team.
Open the assistant
Ask a bounded question
Verify and decide
Boundaries and responsibilities
- Controlled availability. The assistant is available only where it is included and provisioned for the tenant, and access is agreed with Noxtica.
- Assistant text can be wrong or omit context. It is not a verdict, and it is not evidence that a person is fraudulent.
- Risk scores and levels it summarizes are probabilistic and can be wrong. Your team decides what to do with them.
- The assistant does not browse your systems or run your code.
Assistant and connection output is an estimate and can be wrong. Your team reviews it before acting.Read the disclaimer