Botium Box supports all relevant Conversational AI technologies available, and upcoming technologies are added to the platform within weeks. There are more than 30 different Botium Connectors available right now, including adaptable generic connectors for custom Conversational AI APIs.
While it is state-of-the-art for Conversational AI to use some kind of NLP technology in the background, not all those APIs disclose this NLP information to API clients (as Botium Coach). Specifically, the information Botium Coach needs for its analytics:
Predicted intents and entities
Intent and entity confidences
Alternate intent list
The more the API discloses, the better the analytics Botium Coach can provide - you can find details about Botium Connector support here.
To connect your Conversational AI to Botium Coach, click the Chatbots menu and click Create new Chatbot, and enter a name for the chatbot. Click . Select the connector/chatbot technology.
Depending on the choosen technology, you have to provide Botium with endpoint addresses, keys or secrets and additional technology specific information. For example, with Rasa you have to provide:
HTTP(S) endpoint URL of your Rasa chatbot host
Wether you want to use Rasa Core (dialogue engine) or Rasa NLU (choose this for using it with Botium coach) or both.
Authentication or JWT Token (if you secured your Rasa installation)
To check connectivity click Say Hello (Check Connectivity) - this will send a dummy text (“Hello”) to your Conversational AI and print the result. There typically is a welcome message or a greeting intent shown.
Save the chatbot settings and your aready for Step 2: Prepare Datasets for Training and Testing.