How to Teach Anuva
Anuva learns from the dialogues added to its knowledge base. These dialogues are structured in Interests (possible phrases used in user interaction) and Abilities (possible chatbot responses to user interactions).
Since Anuva works reactively to user interactions, in order to form dialogues we need to group interests and abilities in pairs. Contexts will be used only when, during the dialog, it is necessary to store some information that the user refers to for use in another point of the conversation.
It is up to the chatbot administrator to provide this data to the system, structurE the possible dialogues, and carry out the training of the set. The training should be scheduled through the Anuva interface, and at the end of the training, your Anuva will be updated automatically.
Procedure
- Access the menu “Dialogues”;
- Click on "New" and select one of the options: Simple (only an Interest-Ability interaction) or Complex (when joining multiple simple dialogues);
- Inform the Theme related to the Dialogue, this will help you to categorize them, later. A new theme can be registered in the "Theme" menu;
- In the Interests section, complete the phrases related to the user Interest and click on "Save";
- In the Abilities section, complete the fields on the screen, paying attention to the ability type:
- Standard: it will be used to represent abilities that are answered through a predefined text response. Several predefined answer phrases can always be assigned to answer the same question;
- Customized: it will be used when, in order to respond to a user's interest, Anuva needs to seek information on another system. Access the document “How Anuva can interact with other systems” to understand how to use this type of ability;
- Button: it will be used whenever it is necessary to limit user responses to a particular Anuva interaction. When using this option, Anuva will respond to user interaction by displaying buttons. A button is defined by its Name and Value. In the name field, the button that will appear for the user is defined. In the value field, it is defined to Anuva how it should interpret the user's interest by clicking on the button;
- Image: it can be used when you want to respond to the user using text and image. The image associated with the response will be displayed in the chat window.
- Click on "Save".
NOTE:
- Complex dialogues help Anuva to be more assertive in a conversation that is long, use it to guide it to how to respond in those situations; - Anuva's understanding is made by approximation, so the more phrases you add to the interest, the more precise it will be in understanding the user's interaction. Remembering that the user does not have to type the exact phrase that was added to the interest since Anuva works with Artificial Intelligence; - Remember to add at least one dialog with fallback ability. It will be used when your Anuva cannot understand the user's interest or understand with an accuracy of less than 20%. You can view the accuracy of each interaction in the menu Analytics > List conversations.