Test Your Virtual Assistant
with industry-leading NLP+3
Chat with your virtual assistant real-time to test recognition, performance, and flow of the conversation for various scenarios. In-built tools enable you to try different variations of user prompts to ensure the NLP interpreter is processing the synonyms (or lack of synonyms) accurately.
Use this opportunity to improve your virtual assistant performance and responses to customer utterances.
TESTING
Validate Performance
Testing helps you discern & validate the ability of your virtual assistant to correctly recognize the expected intents for a given set of utterances. The Batch testing module allows you to execute a series of tests to get a detailed statistical analysis and gauge the effectiveness of your VA training.
Run automated or customized testing suites to get detailed statistical analysis, record bot user utterances, view NLP metrics like precision, recall & F1 scores.
UTTERANCE TESTING
Interpret Diverse Human Expressions
by testing your assistant with a variety of user inputs
Make sure your bot responds to user utterances with related tasks. Evaluate and test your virtual assistant with a large sample of expected user inputs to obtain insight into bot responses and fine-tune based on the expected response.
BATCH TESTING
Accurately Recognize Intents
by testing the performance of the ML model
Delineate how well the virtual assistant is able to understand user utterances. Using predefined test suites available in the builder or your own custom test suites, execute a series of tests to get a detailed statistical analysis and gauge the performance of the ML model.
CONVERSATION TESTING
Optimize Performance
by testing the recordings
Validate the assistant’s skills end-to-end to ensure the best experience for the customers and improve ROI. The Conversation Testing tool allows you to validate the virtual assistant’s interaction flows across dialog and knowledge graph intents. Record the test cases, execute them, and analyze the test results.
EXPLAINABLE AI
Unlock the Black Box
with a fully transparent NLP engine
The Kore.ai platform provides complete transparency into intent identification and entity extraction with explanations about each query response. Unlock the machine learning “black box” and use actionable insights to perform root cause analysis on various failure scenarios. Use the Platform’s visualization tools, confusion matrix, K-fold testing and NLP analysis to understand the assistant usage patterns, conversation flows and drop-off points. Export and review chat transcripts and other conversation data to further evaluate failure context.
