Setting Up Reported Incident Review Assistant

Overview

The CE Reported Incident Review Assistant is a packaged AI workflow agent that helps agency users review newly reported incidents more efficiently. When an incident is submitted, the assistant analyzes the incident details and generates recommendations to help determine whether:

The selected issue subtype is appropriate.
The incident is a duplicate of an existing incident or case.
The incident priority should be increased based on prior violations or public safety concerns.
Related incidents, cases, permits, or other transactions exist for the same or nearby property.

The assistant displays its recommendations as an AI summary on the Incident Detail page while the incident remains in Reported status. Depending on user security privileges, users may also be able to perform recommended actions directly from the summary, such as rerouting an incident, marking it as a duplicate, or updating its priority.

How the Assistant Is Invoked

The CE Reported Incident Review Assistant is invoked automatically through the Business Rules Framework (BRF) when a reported incident is submitted.

The invocation process is:

A user submits a reported incident.
BRF evaluates the configured business rule.
The Call AI Agent action invokes the published AI agent.
The incident ID and incident key are passed to the assistant.
The assistant retrieves incident details and related operational data.
The assistant performs its analysis and stores the results in the AI results table.
The Incident Detail page retrieves the stored results and displays the AI summary.

Before You Begin

Before configuring the assistant, verify that:

The CE Reported Incident Review Assistant has been copied from the delivered template and published in AI Agent Studio.
The published runnable agent name is available.
The required AI feature is enabled for Code Enforcement.
The Call AI Agent BRF action is available.
You have permission to configure Business Rules Framework.
The system account used by BRF has permission to invoke AI agents.
Incident records contain address or parcel information and latitude/longitude values required for nearby searches.

Enable AI for Code Enforcement

Enable the functional option that allows Code Enforcement to use AI assistants.

The functional option controls whether the Incident Detail page displays the AI-generated summary. If the feature isn't enabled, the assistant won't appear even if the BRF rule is configured correctly.

Review Access Requirements

Review the required security before enabling the assistant.

Verify that:

The BRF runtime can invoke AI agents.
The PCI or system account has permission to invoke the published agent.
Administrators can configure Business Rules Framework.
Users can access the Incident Detail page.
Users who should perform recommended actions have the required privileges.

The AI summary can be displayed to users independently of the action buttons. Action buttons such as Reroute, Mark as Duplicate, and Update Priority are displayed only when the signed-in user has permission to perform those actions.

Configure the Business Rules Framework Rule

Configure a business rule that invokes the assistant whenever a reported incident is submitted.

Navigate to:Navigator > Common Setup > Business Rules Framework

Select the appropriate Code Enforcement offering, and then configure the following sections.

Event: Select the business event that occurs when a reported incident is submitted.

This event should run after the incident has been created and placed in Reported status so the assistant can analyze the submitted incident.
Criteria: Add criteria that determine when the assistant should run.

For example: Incident Status = Reported, Issue Subtype = one or more selected issue subtypes.

Note:
For an initial rollout, Oracle recommends enabling the assistant for one or two issue subtypes before expanding it to additional or all issue subtypes.
Call AI Agent Action: In the Actions section, add the Call AI Agent action.

This action invokes the published CE Reported Incident Review Assistant whenever the configured rule criteria are satisfied.
Family, Product, and Agent Name: Configure the Call AI Agent action using the values from the published AI agent. Specify:
- Family – AI Agent family that contains the published assistant.
- Product – Code Enforcement.
- Agent Name – Published runnable agent name for the CE Reported Incident Review Assistant.
Configure the required input parameters, including:
- Incident ID
- Incident Key
After completing the configuration, save and activate the business rule.

Note:
Ensure to use the exact runnable agent name that was published from AI Agent Studio.

Verify the Assistant on a Reported Incident

After activating the rule:

Submit a new incident.
Wait for the assistant to complete its analysis.
Open the incident from the Worklist or Incident List.
Navigate to the Incident Detail page.
Verify that the AI summary is displayed while the incident is in Reported status.
Confirm that recommendations are generated for issue subtype validation, duplicate detection, priority, and related transactions.
Verify that any available action buttons appear only for users with the appropriate privileges.

Tips and Considerations

Keep the following recommendations in mind when configuring the assistant:

Begin with a pilot rollout for one or two issue subtypes before enabling the assistant for all reported incidents.
Clear incident descriptions improve recommendation quality.
Accurate address and location information are required for duplicate detection and related transaction analysis.
The assistant displays results only while the incident remains in Reported status.
In the current release, editing or reopening an incident doesn't automatically rerun the assistant. Previously generated results may continue to be displayed until a future enhancement introduces on-demand refresh capability.