Extension Flow

This flow captures “prediction of emotion” and “supervisor ask” for every new chat message.

1. When a chat is accepted by an agent, a parent record is created in the ChatAIResultSummary custom object.
2. For each new message in that chat, the emotion model endpoint is called to get emotion prediction.
3. For each new message in that chat, the supervisor ask model endpoint is called to get supervisor ask prediction.
4. Once we have both, the prediction stores the results in the ChatAIPredictionInfo child table.
5. The custom object is updated with the start time and the completed time of a chat, which are retrieved from the product chats table.
6. When the chat has completed, the parent row as is marked as inactive.

The following flowchart represents the extension flow: