Appearance
Verification Policies let admins define which verification questions and rules apply to each department, instead of using a single configuration organization-wide. Departments that handle more sensitive requests - for example, Finance - can be held to a stricter policy than general staff.
Accessing Verification Policies
Go to Settings > Verification Policies to see the list of policies configured for your organization, along with the departments each one is assigned to and when it was last updated.
How Policies Work
- Every organization starts with a Default policy, which cannot be deleted. It guarantees that any department not explicitly assigned to another policy still has a working configuration.
- A policy can be assigned to one or more departments, but each department can only be assigned to a single policy at a time.
Managing Policies
Creating a Policy
Click New Policy, give it a name, and assign the departments it should apply to.
Editing a Policy
Click Edit on a policy to change its name, assigned departments, or configuration (see Policy Settings below). Click Save - a "Changes saved" confirmation appears.
Deleting a Policy
Click Delete policy on the policy you want to remove. A confirmation dialog names the departments it currently applies to: "This policy currently applies to [departments]. Once deleted, users associated with groups will be verified using the Default policy instead. Are you sure?" Choose Cancel or Delete. Once deleted, a "'[Policy]' policy deleted" confirmation appears.
NOTE
The Default policy cannot be deleted.
Policy Settings
Each policy has its own configuration:
| Setting | Description |
|---|---|
| Policy Name | Identifies the policy in the list and when assigning departments. |
| Assigned Departments | The departments this policy applies to. Each department can belong to only one policy. |
| If medium risk detected - apply | What happens when a user in this policy fails verification with medium risk. See Configuring the Medium-Risk Escalation below. |
| Number of verification questions | How many dynamic questions a user must answer as part of verification under this policy. |
| Maximum number of incorrect answers allowed | How many wrong answers are tolerated before verification fails. |
| Require a dynamic question to start verification | When enabled, verification must begin with a dynamic question rather than relying on silent checks alone. |
| Automatic Question Optimization | When enabled, imper.ai automatically adjusts which questions are asked to improve coverage. |
| Questions | The list of verification questions available to the policy, with their data source, coverage, and type. Individual questions can be excluded. |
Configuring the Medium-Risk Escalation
The If medium risk detected - apply setting decides what happens when a user covered by this policy fails verification with medium risk - an uncertain result that isn't safe to pass automatically, but also isn't high-risk enough to block outright. Choose one of three options:
- None - no escalation option is offered. A medium-risk failed verification is treated the same as any other failure.
- Verify with Manager - the agent can confirm the user's identity through the user's manager by email. See Verify with Manager for the full flow.
- Verify via Document - the agent can confirm the user's identity through a government-issued document. See Verify with Document for the full flow.
Because this is set per policy, different departments can be configured differently - for example, field employees could use document verification while other departments verify with a manager instead.
Important Notes
- In this initial release, verification data sources cannot be configured per policy - every policy uses the same organization-wide data sources.