Technology Blogs by Members

Explore a vibrant mix of technical expertise, industry insights, and tech buzz in member blogs covering SAP products, technology, and events. Get in the mix!

1 Introduction

You may have probably already read our Day 1 blog about the first step in moving Skybuffer AI content to the open-source community version. Great, isn’t it? In this case you might have already realized that it is possible to set up automated contact center with seamless integration of a human expert in just 8 minutes by simply forking the AI content from the open-source organization and connecting it to the Hybrid Chats Discovery Cloud.

In other words, you have an out-of-the-box solution with fully operational skills and:

1) the chatbot can greet your clients/leads and maintain a small talk in any language ?

2) the chatbot can validate emails of your potential clients, so you could be sure that the users visiting your website have left the proved email for further communication.

3) you have active Operator Workplace to provide your clients/leads with expert support and enable them to enjoy the integrated communication with both the chatbot and the human support team.

4) from now you will not miss a single request, even when there are no operators online. The chatbot can capture any conversation with the user’s contact details and register a support request in the Hybrid Chats Cloud.

Now we are asking you to kindly spend 3 more minutes and customize your chatbot to provide replies that correspond to the products or services you are selling.

2 Chatbot Customization

2.1 Customize Welcoming Message

Now it is a good time to make your chatbot more suitable for your business goals, for that you can slightly modify the predefined skill.

Step 1: Navigate to the Connect tab of your chatbot.

Step 2: Select webchat or create new one ? (you can always refer to the Day 1 guide to see how to add more webchat channels).

Step 3: Enter a welcoming message that describes and makes it clear for your clients the way this chatbot can help.

Step 4: Save your changes.

3 Customize Chatbot’s Skills

NB! We have a predefined “what can you do?” skill to tell the user about the chatbot’s capabilities. Nevertheless, you can always adjust the reply in that skill, or simply deactivate it.

3.1 Option 1: Skill Deactivation

Step 1: Navigate to the Build section and switch to the List View:

Step 2: Find the get-help skill in the list on the Build tab.

Step 3: Switch off the skill group as it is shown on the screenshot below.

Step 4: In this case, when the user asks about the chatbot’s skills, it will route them to the support team, or will suggest capturing the support request:

3.2 Option 2: Skill Customizing

Step 1: Find get-help-trigger skill in the list on the Build tab.

Step 2: Open the Actions tab of the skill as it is shown below:

Step 3: Find the Action block with the chatbot’s replies in English:

Step 4: Now you can delete the reply we have provided:

Step 5: And you can insert your own reply to “what can you do?” question into the same section:

NB! You can select different types of the message (see the screen below), we will use a simple text in this example:

Step 6: Please, move the reply up, so that it could be placed before the redirection to the skill of zxas-get-help-input:

The result after the drag and drop action:

Step 7: Delete memory parameter rt_skill and redirection to zxas-get-help-input:

Step 8: Done, now you can test you reply in the webchat, or test tool channel:

Step 9: Excellent! And you can always step in as a Hybrid Chats operator:

And here is how the user sees it:

3.3 Add Translation

Step 1: Let us assume your customers speak different languages. In this case you can provide customized replies in any language via the Chatbot Vocabulary application in Hybrid Chats cloud.

Step 2: First let us determine a phrase that is sent to the Chatbot Vocabulary on the SAP Conversational AI platform. We will use the same skill where we have just customized the reply. Find the following Action in the skill with /translate web hook:

Step 3: Adjust rt_source memory parameter with the phrase that you would like to be sent to the Chatbot Vocabulary:

Step 4: Delete rt_list_captions memory parameter from the “Set Memory Field” block:

Step 5: Add memory parameter rt_reply with the value “active”:

Step 6: Save your changes:

Step 7: Delete two text messages blocks after the /translate web hook execution:

Step 8: Delete memory parameter rt_skill and redirection to zxas-get-help-input:

Step 9: Now you can delete all message groups from the skill zxas-get-help-trigger that are following the message group where you have just done the corrections:

Confirm the message group deletion:

Step 10: Move to Hybrid Chats Discovery Cloud to capture the translation in the Chatbot Vocabulary:

NB! In case you do not capture the translation in the vocabulary, /translate web hook can always give you Google translate API key to translate your reply from rt_source in the runtime.

Step 11: In the Chatbot Vocabulary you should provide the skill ID, the target language and the phrase. So, go to the Edit mode:

Click on the Add button:

Capture your translations.

NB! Please use exactly the same phrase you have placed into the rt_source memory parameters and keep JSON formation that is set by SAP Conversational AI (quotes in our case).

After you save it, the chatbot will take the phrase from the Vocabulary.

3.4 Clean up the zxas-get-help-floating Skill Group

In the modified “get-help” skill we do not expect any input on the user's part, so the skill of zxas-get-help-input can be deleted to clean up the zxas-get-help-floating skills group.