How to use ChatGPT to analyse workflow conversations

This session will show how ChatGPT can categorize and summarise workflow conversations.

Pyramid allows users to start a dialogue against a particular data point. I have added some comments to a conversation workflow and assigned it to a colleague or a group. The conversation flow and history can be viewed in Pyramid or other collaboration platforms.

All conversation data can be directly queried and shared with business users.

Here, we are using text aggregation to consolidate conversations per data model and then applying ChatGPT to summarise key insights from the conversation history.

I can view the number of conversations by data point by selecting a data model. I can also see the conversation history by date and time.

Here, we have used the text aggregation measure to view the entire conversation in a single sell.

By selecting one of these conversations, Pyramid will use our OpenAI integration to provide insights into the data. OpenAI will respond with three aspects. The first shows the main subject categories discussed in the conversation. The second is a conversation summary, and the third classifies if this conversation led to a resolution to the problem or not.

In this session, we have shown how Pyramid can uniquely track workflow conversations by data point, consolidate the text using its text aggregation feature, and then apply ChatGPT to provide insights.