Computing performance scores of conversational artificial intelligence agents

摘要

Disclosed herein are systems and methods of generating a score for artificial intelligence chatbots. In some embodiments, the method includes receiving configuration data that is pre-configured as well as receiving input data which can be in many forms including historical chatbot conversation logs, real-time chatbot conversation data, or simulated chatbot conversation logs based on simulated users generating prompts via artificial intelligence. The input data is then parsed based on the configuration data and then processed. The parsed data processing can include generating a variety of scores including a navigation efficiency score, a compliance score, an intent score; and a sentiment score, among others. Upon processing, an overall score can be generated based on at least the configuration data, the navigation efficiency score, the compliance score, the intent score, and the sentiment score. These overall scores can be utilized to generate further correlation reports, and can be configured to specific implementations.