Co-Pilot in Data Intelligence

In This Article:

• DvSum configuration
• 1. Add Source
• 2. Configure New Chatbot
• 3. Analyzing Customer Data from Train Bot
• How Chat History Works with Copilot

This article describes the steps needed to configure Rest API as a source in DvSum Data Intelligence (DI) to use CADDI Chatbots. 

DvSum configuration:

1. Add Source

To create a data source, navigate to Administration → Data Sources → ⊕ Add Source.

Select Rest API.

Give the source a name, and save it.

2. Configure New Chatbot

Click on "New Chatbot" which will redirect the user to the new chatbot detail page.

From the Definition tab click on the "Download sample file" button to download the sample file. Configure the workflow JSON according to the requirements and upload the file.

There will be a default prompt already being set but the user can also update the prompt of the bot accordingly. The prompt directly impacts the quality of the results.

Once the workflow JSON file is uploaded and the prompt is updated, save the chatbot.

3. Analyzing Customer data from the Train bot

Navigate to the Train bot tab and type a question containing the customer ID.
When the question is executed, Co-Pilot displays a progress indicator showing the following four steps

1. Understanding your question – Co-Pilot interprets the query, identifies parameters, and validates inputs.

2. Selecting relevant tool – The most appropriate analysis tool is automatically chosen based on the question type and detected parameters.

3. Fetching data – Relevant records or data sets are retrieved for processing.

4. Analyzing data – The data is examined to extract key insights and generate a summarized result.

Once all four steps are completed, insights are displayed automatically, giving you a clear summary of findings and any key observations derived from the data.

The results display key facts, observations, and recommendations derived from the data. The prompt set from the Definition tab influences these results.

Insights from the data would show up in different grids that can be analyzed.

How Chat History Works with Copilot:

Users can ask a variety of questions related to telecom troubleshooting beyond analyzing customer IDs. Copilot utilizes a Q&A chain specifically designed for telecom troubleshooting queries. Users can also ask follow-up questions to receive recommendations or additional troubleshooting assistance for specific customer IDs.