Test and Debug Overview - NLP

Back to NLP Topics

After building and training your app, use the platform’s testing suite to validate behavior before publishing.

---

Testing Tools

Tool	Purpose
Test (Talk to Bot)	Chat interface to test conversations in real time. Validates dialog flow, NLP, and task execution.
Utterance Testing	Enter utterances and see which engine matches them and which intent wins. Identifies ambiguity and incorrect matches.
Health and Monitoring	Analyzes training data, test coverage, and results. Identifies intent issues and provides training recommendations.
Batch Testing	Runs a set of utterances against the ML model to get statistical accuracy metrics (F1, precision, recall).
Conversation Testing	Simulates end-to-end conversation flows. Create test cases for business scenarios and run them for regression.