Overview

The logging functionality is an essential module in GPTBots for recording and managing user-Bot conversation logs, Q&A, and actions related to question categorization. It supports features such as training improvement, quality inspection, sentiment recognition, and question categorization based on conversation logs.

Training Improvement

For AI Bot response logs, administrators can perform training operations and add modified logs to the knowledge base dataset to optimize the response quality of the AI Bot. Changes take effect immediately after training.

Quality Inspection

Enterprise operations managers can review conversation logs and perform quality inspection operations on the AI Bot's responses. The quality inspection results are categorized as: Resolved, Unresolved, and Partially Resolved.

Sentiment Recognition

When the "Question Categorization" feature is enabled, the system automatically performs sentiment recognition on user queries in the conversation logs, labeling them as Very Positive, Positive, Neutral, Negative, and Very Negative.

Question Categorization

When the "Question Categorization" feature is enabled, the system identifies and matches user queries in the conversation logs with enterprise-defined question categories and tags each user query with its corresponding category.