Connecting Table Data to an Autonomous Node in Botpress

Autonomous nodes in Botpress can seamlessly integrate with structured table data to answer questions, recommend products, and maintain controlled interactions. This lesson covers the process of connecting table data, configuring queries, and fine-tuning the bot's behavior for a reliable user experience.

Connecting Table Data to the Autonomous Node

1. Adding the Query Knowledge Base Card
   • Open the autonomous node and add a "Query Knowledge Base" card from the actions menu.
   • This card includes two inputs:
     • Query: The question the bot will use to retrieve data.
     • Knowledge Base: The specific knowledge base or table to query.
2. Importing Data
   • Create a table in Botpress, such as "Tractors Table."
   • Import data from a CSV file containing relevant fields (e.g., name, description, horsepower, and price).
   • Ensure all fields are searchable to improve the bot’s ability to provide accurate answers.
3. Linking the Table to the Knowledge Base
   • In the knowledge base settings, add the imported table to make its data accessible to the bot.
   • Configure search fields, such as name, description, power, and price, to refine the bot's responses.

Defining Objectives and Behavior

1. Adding Context in Instructions
   • Use the instructions section to define the bot’s objectives. For example:
     • Answer questions about BP Tractor products using the knowledge base.
     • Ask clarifying questions to understand user needs.
     • Recommend up to three options at a time to avoid overwhelming the user.
2. Testing and Adjusting
   • Interact with the bot in the emulator to verify it generates accurate responses.
   • Adjust the number of options or other behaviors based on user feedback and testing.

Controlling and Scoping AI Behavior

1. Preventing Misleading Responses
   • Define specific behaviors in the instructions to avoid unintended outputs. For instance:
     • Specify that prices in the knowledge base are final and include all discounts to prevent the bot from fabricating discounts.
2. Example Scenarios
   • Without proper instructions, the bot might incorrectly provide a discounted price when prompted.
   • After updating the behavior settings, the bot can respond appropriately, ensuring users receive consistent and accurate information.