AI that works
within your rules
Role-based permissions, AI guardrails, and customer data protection—run support automation by your organization's standards.
Permissions and roles
| Role | Chats | Settings | Billing |
|---|---|---|---|
| Owner | |||
| Admin | – | ||
| Member | – | – | |
| Billing admin | – | – |
Permissions:
access by role
Split roles into Owner, Admin, Member, and Billing admin, and control which screens each can see and which settings each can change.
Invite teammates by email and assign their role as you invite them.
Separate teammates who handle conversations from those who only manage operations, so conversations go only to the right people.
AI guardrails
Block answers beyond knowledge
Hand off sensitive requests
- 2 conditions
Refund requests
Eligible orders only
Execute cancellations
Guardrails:
lines AI won't cross
LIA never makes up anything beyond its trained knowledge and connected data.
Requests outside your policy, or sensitive ones, are logged and recorded by the rules you set.
LIA Actions run only within the actions and conditions your team allows.
Customer
My number is (555) •••-••••
Data protection:
customer data belongs to customers
Conversations and customer data are encrypted both in transit and at rest.
Sign in with your organization account and limit console access to approved networks.
We guide you at upload time so sensitive data stays out of training material.
Where your rules apply
What actions can do, who sees analytics, and what knowledge is used—all follow the same rules.