May 2023
NLU Fallback Model Improvement & NLP Scoring
In this release, we're excited to introduce new capabilities aimed at enhancing query handling and benchmarking the performance of our Natural Language Understanding (NLU) engine over time. These enhancements represent our commitment to continuously improving the reliability and effectiveness of our conversational AI platform
Highlights:
-
LLM Query Resolution : We've integrated and bundled our NLU with a Language Model (LLM) into our system to handle queries that our bot is unable to answer. When encountering an unanswered query, the LLM steps in to provide relevant responses, ensuring a seamless user experience even in complex scenarios.
-
NLU Performance Benchmarking : Introducing a comprehensive benchmarking framework to evaluate and monitor the performance of our NLU engine over time. Tracking key performance metrics, such as intent recognition accuracy, entity extraction rate, and dialogue success rate, and compare them against historical data to identify trends and areas for improvement.
