How can I configure the Proofread Agent feature for BUI Incident Threads?
Environment
Oracle B2C Service Browser UI
Requirements:
- An AI Service to provide the proofreading functionality
- External Objects is enabled for your site
- This feature does not require a CX upgrade
- This feature does not require a feature flag
Resolution
To improve the experience for agents, we've introduced the Proofread Agent for incident threads, an AI-powered service that reviews your messages before they are sent to customer.
In the past, agents could only use basic browser-based spell checking, with no dedicated proofreading tool available. With this update, the Proofread Agent is now built directly into the incident response and private notes thread, making it easier to review and refine your messages for accuracy and professionalism before sending them to customers.
Details of the Agent Experience
- The Proofread Agent feature must be enabled for the agent’s profile
On the Response or Private Note thread type, a new icon will display in each thread header of an unsaved thread. When the agent clicks on the icon, the content of the thread will be sent to your AI service of choice.
By default, the icons for copy and replace will display on the dialog.
Copy - The agent can copy the content and paste it into a thread.
Replace - The agent can click replace and the content will automatically be updated in the thread.
The Feedback for thumbs up (good) and thumbs down (bad) requires customization using extensions by the administration.
Administrator Experience:
Configuration Details:
Step 1 - Create the integration within External Objects.
An External Objects Connection is required for connecting the Incident Thread Proofread feature with sources like OCI Generative AI or Other REST endpoints.
To set up External Objects follow these steps.
- Within the Browser UI, Navigate to Administration > External Objects & Integrations
- External Object Home Page will list all existing connections
- To configure the Proofread Agent endpoint, create a new External Object connection. If you plan to test multiple AI service providers with this feature (e.g., OCI Generative AI, Azure Open AI), each provider requires a unique endpoint. Therefore, you must create a separate External Object connection for each AI service.
- Select "Add New" for creating a new External Object Connection. Provide the API endpoints of the service provider and include the necessary authentication details. Ensure that you add the required headers or query parameters as specified by your service provider.
For more information on External Objects please reference this documentation.
Step 2 - Create a new connection for the Proofread Agent feature
- In the Administration page in BUI, navigate to AI Studios > Proofread Agent for Incident Threads.
- The first tab you will see a list that will contain any configurations already set up. If this is your first-time configuration Proofread, this list will be blank.
- Select Create New and populate the following fields.
General Settings
| Field | Description |
| Name | A descriptive name for Proofread Agent connection |
| Provider | Select the AI service provider. Options include: OCI Gen AI Agent, OCI Gen AI Chat, Azure Open AI, or Other |
| Connection | Select the integration created in External Objects |
| Maximum Character Limit | Defines the max allowed character count for user questions |
| Created | This auto-populated field displays the time when the Proofread connection was created |
| Created By | This auto-populated field displays the time when the Proofread Agent connection was created |
| Updated | This auto-populated field displays the time when the Proofread Agent connection was created |
| Updated By | This auto-populated field displays the time when the Proofread Agent connection was created |
| Notes | Enter any relevant notes for the Proofread Agent connection here |
Profile Access Configuration
- Select which agent profiles are granted access to the configured Proofread Agent.
- Note: A single profile cannot have more than one Proofread Agent configuration per interface.
Extensibility Options
| Action | Description |
| Enrich Question Extensibility Action | Modify the question before sending it to the AI service |
| Enrich Answer Extensibility Action | Modify the AI response before displaying it to the agent |
| Logging Extensibility Action | Track usage by logging interactions in a custom object or table |
Example Usage:
- Create a custom object to log Proofread Agent interactions.
- Use custom reports for analytics and auditing.
