Goal | Metric |
---|---|
Dealing with unstructured input | Successfully answering queries without triggering Fallback Intents. |
Engaging Conversational experience | Check the Session Flow and see where the users are exiting a chat (and whether they’re exiting the chat as intended.) |
Understand what users mostly talk about | Through analysing sessions and interactions. Sentiment Analysis. Text Analysis (next steps). |
Provide quick support to clients | Enabling ‘Community Support’ and monitoring the Forum. |
Automating FAQ Generation | Decreases the number of intents created and handles static questions through Knowledge Base. |
Decrease time taken to configure the Bot | Paraphrasing ‘Training phrases’ with ML Models. Compare manual time taken to write custom training phrases and ML Model’s automatically generated training phrases. |
Able to handle small talk and tell jokes. | Check the feedback from users and their queries while interacting with the bot and see if it meets our needs. |
Multi-Language Support | Train the bot in a language other than English and check if it’s up to the mark. |