Creating QnAs
A QnA is a question-answer pair. The answer responds to a simple question.
The QnA is a special type of conversational flow designed to simplify the business user experience when creating simple conversations. They are comprised of training phrases (primary and alternate questions) and a single step of type Message that delivers the response to the user (answer).
This section describes how to create KB articles.
Prerequisite
- You need to have Knowledgebase Author security role, otherwise you will not see the QnA in solutions. If you don’t have this security role, get in touch with your DRUID Portal administrator.
Step 1. Test your primary question
To test your primary question, access the QnA. In the User says field, enter your primary question.
Click the icon and select the desired bot language.
DRUID searches the question against all intents that exist on the bot (in the selected bot language). If the primary questions is found on a QnA or flow, that means that the chatbot already knows how to answer that question. DRUID displays the search result(s) in either QnA OR Flows based on where DRUID found the intent.
If the primary questions is found as intent on a flow(s) you see that it exists but as Knowledgebase Author you do not have access to modify the identified flow(s).
A primary question passes the test when DRUID finds no matches. Using this primary question, you can now create a QnA.
Step 2. Create QnA
At the right-top side of the Questions and Answers page, click the Create New QnA button ( ). The Create QnA page appears.
Provide the QnA details described in the table below.
Field | Description |
---|---|
Name | The name of the QnA the DRUID platform uses in the backend . The field is mandatory. |
Description | The QnA description. This field is mandatory.
You can localize this field in the bot languages. Provide description in the bot languages displayed on the page. Note: If the bot understands more knowledgebase articles (QnAs) as being appropriate for the user question (intent), the bot will display the description associated with the user’s question and give the user the possibility to choose the KB article. Important! If you add the primary question in the Description field, you should add it in the Training Phrases as well. Example: For DRUID version 1.55 or higher, Authoring time translation is available for QnA description (if set on the bot). To see QnA description in a specific language, select it from the Languages field. |
Category |
For easier filtering and increased efficiency, DRUID creates QnAs and breaks them down into categories. Provide the desired category for your QnA. This field is mandatory. |
Response Message |
The answer to the user’s question. The message the DRUID bot will respond to the user. The response is mandatory. Provide the answer in all bot languages displayed on the page. Example: For DRUID version 1.55 or higher, Authoring time translation is available for QnA response message (if set on the bot). To see QnA response in a specific language, select it from the Languages field. |
Training Phrases |
In this section, you add the primary questions and the desired alternate questions (if any). The questions are localizable, so add them in all bot languages displayed on the screen. Enter question and click the add icon displayed inline ( Note: The first question you add in the Training phrases section is marked as being the primary question; therefore, make sure that you first add the primary question and then subsequent alternate questions.
Example: You can delete alternate questions by clicking the delete icon displayed inline. For consistency purposes, when deleting alternate questions, make sure to delete it in all bot languages. |
Requires Authentication | Select Authentication Required only if the QnA requires authentication. Useful if the bot is talking to existing customers or employees. |
Click the Save button to save the KB article.
The Article appears under the specified category.

Step 3. Train the bot
To include all questions defined on the QnA flows associated with the respective bot training model, you need to train the bot. For that, go to the bot details page, click the NLP section header, then click the Train button. The bot NLP model will be trained to understand the questions as well.