Finish here: How to connect the bot to the extension

Now you created questions and answers in your knowledge base using the QnA Maker Extension, you need to link the bot and said base. To do it, follow the steps below and execute them only once per bot.

1. Access the Builder
2. Select the box that represents the flow point that will respond to the user’s input/question
3. When you click on the box, a menu will be opened to the right. Select the “Actions” option of this menu:

 

4. In “Actions”, click on “Add input action” and select the option “HTTP request”

4. To do the HTTP request, select POST as “Method”

4. Then, in each field, provide the corresponding data, as instructed below:
5. Fill the URL field with the following link:

https://microsoft-qna-maker.cs.blip.ai/api/QnAMaker/v1/knowledgebase/question

 

Result:

 

2. There are two fields in “Header.”
3. Fill the first field, “Key,” with the following word:

BotId

 

2. Fill the second field, “Value,” with the ID of the bot you are installing in the QnA Maker extension.
3. To get this piece of information, right-click on “Home,” in the top left corner sub-menu, to open the bot’s home in a new tab of your browser:

2. In this new open tab, in the bot’s Home page, look for its ID, right under its name. Copy it.

3. Then return to the previous tab, with the Builder open, to complete the HTTP request. Paste the bot’s ID in the “Value” field, on the header.

 

Result:

 

3. In “Body,” delete any preexisting text and paste the information below:

{

  "question": "{{input.content}}",

  "top": 3,

  "isTest": true,

  "scoreThreshold": 50

}

 

 

Result:

 

7) In “Save Responses,” fill two fields:

1. a) Fill the first field, “Variable for the response’s status,” with the following word:

statusQnA

 

1. b) Fill the second field, “Variable for the response’s body,” with the following word:

responseQna

 

Then, we will have the result:

Done! You completed your HTTP request, and this should be the full result:

8) Now you need to format the request content that will be displayed to the user. To do it, you must:

1. Click on “Add input action” and select “Run script:”

 

You will find it under the HTTP request, as in the print above.

2. In “Run script”, click on “Input variables” and add the variable created in the request:

responseQna

3. Now click on Script

 

4. In the opening box, delete the preexisting text and paste the text below:

 

function run(input) {




    let parse = JSON.parse(input);




    let response = parse[0]







    return response;

}

 

This must be result:

5. In the right menu, in the field “Save return,” write your script’s return variable name. In this case:

 responseFormated

 

 

9) Now, to finish the process, you must make the bot use the request created to respond the user. To do it, you must:

 

1. Access the right menu’s “Content” area, in the same box you used to create the request:

1. In the flow point of your choice, click on the

icon and add a text-type input clicking on

1. Then click on

to edit the field, delete preexisting text, and fill it with the following text:

{{responseFormated}}

 

This will be the result:

 

Done!

Once you complete these steps, the bot will be linked to the knowledge base in the QnA Maker, using this Artificial Intelligence in the bot’s and in the flow’s assigned point. In turn, it can be edited via the QnA Maker Extension.

 

 

 

Attention point: CREATE, EDIT, AND TEST!

 

With the bot fully linked, you can directly test the QnA Maker’s knowledge base you are creating or editing. Just click on the “Test” icon, on top right corner, and on the previously established flow point, where the HTTP request was made, as described in the steps above, and the QnA Maker’s AI will be used to respond to the smart contact user’s inputs and questions.

For more information, access the discussion about this topic in our community or watch our channel’s videos. 😃