Environment:
Oracle B2C Service BUI

Resolution:

Hot Topics is an Embedded AI service within Oracle B2C Service designed to provide administrators with insights into the most common and emerging issues within their incident data. By analyzing incidents based on contextual similarity, Hot Topics groups incidents that share similar characteristics and assigns relevant topic names to these groups. This feature helps admins to quickly identify the trends that are contributing the most to incidents in the system, enabling them to focus on resolving the issues that matters the most.

Hot Topics offers administrators the ability to track major issues, identify new topics that may be forming, and gain a deeper understanding of customer concerns based on incident trends.

Being an Embedded AI service, Hot Topics includes pre-configured AI pipelines and is fully integrated with Oracle B2C Service, requiring no additional setup for the core AI components. This ensures a seamless user experience where administrators can begin utilizing the service immediately once it has been provisioned for their site.

How to Enable and Provision Embedded AI Services or Hot Topics

To get started with Hot Topics, your Oracle B2C Service site must be running CX version 24B or higher. If your site is in one of latest versions, you can proceed by raising a Service Request (SR) to enable the Embedded AI Services tile in the BUI Administration Panel.

Once the tile is enabled, administrators can begin provisioning Hot Topics by following a series of steps within the Administrator UI.

1. Navigate to the Administrator UI in the agent browser.
2. Open the Embedded AI Services tile.
3. Acknowledge the popup that outlines the AI provisioning process by clicking "Ok."
4. Read and Accept the terms related to data usage by checking the Acknowledge checkbox and click Proceed.

Once provisioning is complete, which can take several hours, administrators can access the Hot Topics tile and view the Hot Topics report, which provides insights into the top trends emerging from the incidents in their system.

How Hot Topics Works

The AI models behind Hot Topics analyze incident data to group incidents based on contextual similarity. This means that incidents that share similar context based on subject and first thread are grouped together into “topics.” Each of these groups is assigned a relevant name, providing context to the issues that are being raised. The trends generated by these insights help administrators understand which problems are most common, identify new or emerging issues, and prioritize actions to address the most pressing concerns.

The AI models work by processing data through machine learning techniques to identify patterns and similarities in incidents. The results are then presented in the Hot Topics report, where administrators can view a list of top incident groups based on frequency and importance. This can help identify systemic issues, areas requiring attention, or even emerging trends that might indicate new or growing problems within the system.

Data Usage and Security

The data processed as part of the Hot Topics feature is stored and maintained in isolation for each tenant and site. It is retained for up to six months and can be removed at any time if the customer disables or opts out of the feature. This data is used solely for configuring Hot Topics and, by extension, any future features opted by the customer that utilize Hot Topics as the underlying model within the specific site. It is never shared across tenants, sites, or external systems. Even for sites owned by the same company or customer, data remains strictly segregated, and the same topics are not expected to be available in test sites. Additionally, all personally identifiable information (PII) is removed before data is processed, ensuring compliance with data privacy and security standards.

Initially, the data source will handle up to 100,000 records from 1 month of incident data during the initial ingestion process. After provisioning, the service requires near real-time incremental data updates to reflect any new incidents and ensure that the latest trends are captured (A delay of couple of minutes is expected). This means that new incident data will continue to flow into the system, keeping the Hot Topics report up to date.

Limitations of Hot Topics

While Hot Topics provides valuable insights, there are some limitations to be aware of:

1. The initial data source (one time data tagging) can only handle up to 100,000 records from a maximum of 1 month of incident data.
2. After the initial data ingestion or one time tagging, the system requires near real-time updates to capture and display the latest changes in incident trends. This means that it is important to continue feeding new data into the system for accurate mapping of to.
3. The total number of records that can be processed by Hot Topics is capped at 700,000 records for up to 6 months of incident data. This ensures that the service remains efficient and does not get overloaded with data.

Hot Topics is a powerful tool for administrators using Oracle B2C Service, providing valuable insights into the most frequent and emerging issues within incident data. By leveraging AI to group incidents based on contextual similarity, Hot Topics helps identify problem areas, trends, and new topics that may require attention. By enabling and provisioning this feature, administrators can gain a deeper understanding of the issues affecting their system and take proactive measures to resolve them.

Note: When enabling this feature for a test environment, ensure that the site contains a minimum of 1,000 resolved incidents to support successful provisioning.