Skip to Content
Technical Articles
Author's profile photo Siarhei Tsikhanenka

Day 1 | Skybuffer Enterprise-Level Conversational AI Content Made Public for SAP Community Development

1. Mission, Vision and Value

1.1      Mission

Since July 2018, our company has been deeply engaged with onboarding SAP Conversational AI solution as integral strategic part of SAP Cloud software family offering and has been building expertise around it.

There have been lots of positive, negative and, quite often tough moments, we have faced loads of happiness, progress and failures, too – all those stages that any team comes through while building a road of expertise in the innovative domain moving mountains to reach the point of enterprise-ready solution.

Our team have talked about SAP Conversational AI to more than 400 companies, made over 100 demos and pilot projects converting some of the audience into happy customers. Also, we have widely talked about SAP Conversational AI and our experience at webinars organized by SAP or SAP User Groups. Thank you all for your valuable contributions of your business cases, questions, concerns that have made overwhelming impact on our understanding of enterprise needs out of SAP Conversational AI. The point to be made here is that we deliberately do not use the term of chatbot in order to draw a clear line and distinguish between “see it is easy, I can do the same in a few hours” solutions and solutions that are enterprise-ready.

30 months of development is a great milestone to step up and try to introduce some changes into the game that all of us are playing in the SAP Conversational AI community, and some of us in enterprise tenants. We made a crucial decision to really boost SAP Conversational AI global expertise of SAP Partners, SAP Clients, SAP Leads and, I am sure that it is a right assumption, SAP Internal Development team as a whole.

Right now, we are opening a source of SAP CAI foundation part of the solution, because it is a key part of the overall solution, and SAP CAI content is completely developed on our own, and not limited by any third-party licenses.

We also plan to open node.js engine for the bot management (currently available as a “discover” version on our portal), however we need to review and adapt the code that is based on the third-party node.js libraries.

1.2      Vision

We would like to unite SAP Partners, Clients and SAP Development team around the open source SAP Conversational AI content we have developed and shared in the community tenant. We see Skybuffer as a “moderator” of open source AI content and AI content development methodology that allows developing enterprise-ready chatbot solutions in the low-risk and cost-efficient manner utilizing object-oriented, conversation-driven content development approaches.

1.3      Values

Providing SAP Conversational AI content, we help companies to immediately onboard:

  • Multilingual chatbot development content that is based on both translation APIs and vocabulary of the chatbot’s replies
  • Object-oriented skills development methodology that ensures easy scalability of the chatbots you develop
  • Conversation-driven content development that makes chatbot’s behavior smart and closer to the expectations of the customers, especially in B2C scenarios
  • Operator workplaces (as part of Hybrid Chats Cloud offering from Skybuffer, that is by nature multitenant SAP Partner Managed Cloud, means SAP software family in any case).
  • Automatic categorization of all conversations up to 4 levels of hierarchy
  • Categorization-based authentication to route conversations to agents in a more efficient and secure way
  • Business analytics of conversations contributing to getting more insights for new skills development and understanding gaps in the content that can be covered by appropriate training.

Let us start getting mutual benefits and building community following the “Day 1” guide. We will post more information on weekly basis explaining the object-oriented AI content development methodology in detail and, hopefully, onboarding contributors that can also support us in our Mission.

And yes, we are selling Hybrid Chats solution for SAP Conversational AI, but those posts are not about it. Naturally, like any company or community, we need fuel to survive and move ahead, so we opted for Hybrid Chats solution that brings additional benefits to organizations support services, however, we do not have mission to advertise it in detail in SAP blogs, we have other channels for that. Let us focus on SAP Conversational AI content, features and expertise development. Hope you do understand us and are ready to join!

2. Create User and Organization on SAP CAI Platform

NB! If you already have a user ID and organization created on SAP CAI Platform, skip this step.

2.1      Create SAP CAI User ID

Step 1: Open https://cai.tools.sap/ and Sign up

 

Step 2: Enter the Credentials, agree with Terms and Conditions and click the Register button.

2.2      Create Organization

Step 1: Login to your account on https://cai.tools.sap/

Step 2: Click the button of Create an organization under your profile picture

Step 3: Enter your Organization name and provide its description (optional). Select Private type of new organization. Click the button of Create organization.

3. Fork Skybuffer Community Chatbot to Your Organization

Step 1: Open the Foundation Content chatbot (public organization managed by Skybuffer).

Organization: https://cai.tools.sap/skybuffer-community

Chatbot: https://cai.tools.sap/skybuffer-community/skybuffer-foundation-content

Step 2: Click the Fork button in the top right-hand corner of the page

Step 3: Select your organization from the drop-down list on the pop-up screen.

Step 4: Click the Fork button.

Step 5: Done. The chatbot will be forked to your organization in a few minutes. You will receive an email notification from SAP CAI Platform when your bot is ready.

Welcome to Skybuffer Community 🙂

The initial version of our chatbot has the following major foundation skills:

  • floating skills wrapped with object-oriented AI development methodology
  • business partner creation for leads registration
  • support requests capturing
  • email validation with one-time code
  • phone validation with SMS (Twilio tenant connection required)

All the skills have /translate web hook implemented that can either translate in the runtime or get a reply from the vocabulary, if it is defined there. Vocabulary is a mapping Fiori application that allows the chatbot to reply in various languages thanks to the inner bot management engine.

All the skills are developed with a business fallback option that allows going back to the skill based on the memory configuration (runtime conversation context).

All the skills have /adjust_memory web hook implemented that allows performing sentiment analysis for any language the user speaks to the chatbot.

In general, it is a fully functional chatbot of the entry level that helps to capture information on leads, create a support request or seamlessly integrate human expertise in the live mode via Hybrid Chats. You can immediately use it on your website guided by operator in the first weeks while you are adding more custom-specific skills in order to increase the level of cognitive automation.

4.      Bot Management Apps

4.1       Register your Intelligent Assistant

Step 1: Open the newly-forked bot in your Organization.

Step 2: Go to the chatbot settings to configure the bot.

Step 3: Here you can change your bot’s name, adjust the Data Policy, and set up specific Permissions for your chatbot.

Step 4: Go to https://discover.skybuffer.com/ to access the Bot Management Apps.

NB! you can use administer user ID for our discovery cloud organization:

Discovery User: HC_DEMO

Password: Demo123

Step 5: Open the app of My Intelligent Assistant.

Step 6: Click the plus button to add a new bot.

Step 7: Fill in the data fields on the screen.

Fields description. Where to find them?

  • Intelligent Assistant ID – the Chatbot’s Name
  • Intelligent Assistant Description – your bot’s description
  • SAP CAI User ID/Organization – name of your organization created in step 2.2
  • Version – Active version of your bot (v.1 by default)
  • Active Switch – activate your bot’s settings
  • Enterprise Tenant – Name of your company (not modifiable/configured in the backend)
  • Middleware RFC – RFC connection to DB and servers. Select the appropriate one from the drop-down list.
  • SAP CAI AUTHENTIFICATION TAB: Developer Token – Token to access your bot.
    Copy it from the bot’s settings (see screenshot below) and paste in the app.

  • Google API Key – Add your Google key and enjoy automatically translated communication (without it the bot will only use the English language by default)

NB! As soon as you enter all the data, SAP CAI Bot Webhook Base URL and Hybrid Chats Connection URL will be completed. They will be used in the Bot Settings configuration.

  • Language Control – the chatbot speaks all languages by default (if you have Google API key), but you can select the set of languages (forbidden or allowed) and set up a default language for your chatbot. If you add a language to the Forbidden tab, it means that he chatbot can’t use this language in its replies, instead of the indicated forbidden language the chatbot will be use the Default Language. If some languages are added to the Allowed tab, the chatbot will only communicate in these languages and use Default for others.

Step 8: Save your bot.

Step 9: Connect to the Hybrid Chats backend to get multi-language functionality and communication channels maintenance functionality be activated.

  • Get SAP CAI Bot Webhook Base URL from the Bot Management application
  • Paste it into the Bot Builder section of Version configuration in Settings

Step 10: Connect the bot to the Hybrid Chats front-end solution to get messages from WebChat channel into Hybrid Chats – Operator Workplace.

NB! please note that messages from TestTool are not transferred to Hybrid Chats – Operator Workplace.

  • Copy the link for the Hybrid chats connection link field.
  • Paste it into my Bot Connector sections in the Settings tab.

5. Create Web Chat Connector

Step 1: In your bot, go to the tab of Connect.

Step 2: Add a new webchat

Step 3: Customize your webchat.

Step 3.1: Create your webchat name.

Step 3.2: Select the color scheme for the chat.

Step 3.3: Customize the Chat Header.

Step 3.4: Set up Messages, Bot and User Pictures.

Step 3.5: Select the chat logo and Click the Title message.

Step 3.6: Set up the Conversation saving time and other parameters in the chat.

Step 3.7: Create your new channel. Webchat script should be generated.

6.   Register Web Chat and Test Tool Chanel

Step 1: Go back to Skybuffer Discovery Cloud https://discover.skybuffer.com, open the channel management application.

Step 2: Register a channel for the Test Tool

NB! Channel registration is necessary to simplify the initial chatbot memory setup. We will describe in the following blog posts what types of memory parameters we use in the AI content development methodology and how those parameters influence the chatbot’s behavior in different channels.

Step 3: Register Web Chat channel

Step 4: As soon as you complete channels configuration, it is necessary to move those settings to node.js engine for the chatbot management (the application is also a candidate to be moved to the open source environment in 2021). To do this, you need to open My Chatbot application and click the push button on the chatbot details screen:

Step 5: Now your chatbot has the multilingual feature activated with a possibility to take responses from the chatbot’s vocabulary. Additionally, conversations automatic categorization, live chat with operator, business analytics and many more nice features are available out-of-the-box.

Step 6: Enjoy chatting with your bot in any language!

Step 7: Or communicating back to the user as the Operator from the Discovery Skybuffer Cloud.

Conclusion

So, you see that just 8 minutes of using Skybuffer open-source AI content, and also Hybrid Chats solution, make it possible to set up your support organization, for example, directly from your website or Facebook messenger, and get visitors’ contact data capturing their names, emails or phone numbers, provide expertise via operator workplace or capture a categorized support request in case operators are not online.

Yes, that is just a first tiny step, still, it is enterprise-level automation that can grow and scale easily by getting and analyzing information from your conversations, deploying new skills, or involving necessary human expert.

We will keep posting blogs and adding more skills to the community AI content on the weekly basis, so you can keep following us and grow either your expertise or your support automation solution.

Enjoy SAP Conversational AI – a true no-code flexible platform that allows building context-driven conversations, implementing object-oriented skills building methodology that we will also describe in our future posts.

Hope you enjoyed this reading and the time you spent!

P.S. You can also find the entire list of our blog posts under the links below:

Day 1 | Skybuffer Enterprise-Level Conversational AI Content Made Public for SAP Community Development

Day 2 | Skybuffer Community Chatbot | How to Customize Your Chatbot “get-help” skill

Day 3 | Skybuffer Community Chatbot | How to Create Multilingual Question-and-Answers Scenario Fast and Easy

Day 4 | Skybuffer Community Chatbot | How to Bring Your Own Bot to Hybrid Chats

Day 5 | Skybuffer Community Chatbot | How to Create Multilingual “Question and Answer” Scenario

Day 6 | Success Story | SAP Innovation Award | Cognitively Automated Customer Care

Assigned tags

      4 Comments
      You must be Logged on to comment or reply to a post.
      Author's profile photo Paul PINARD
      Paul PINARD

      Awesome blog Siarhei Tsikhanenka!
      Looking forward to the next ones! 🙂

      Author's profile photo Siarhei Tsikhanenka
      Siarhei Tsikhanenka
      Blog Post Author

      Thank you, Paul PINARD ! Day 2 is posted :), working on the next posts.

      Author's profile photo Somnath Paul
      Somnath Paul

      Quite a useful and started following. Thank you!

      Author's profile photo Siarhei Tsikhanenka
      Siarhei Tsikhanenka
      Blog Post Author
      Thank you, Somnath!
      Don't hesitate to come back with questions, we are quite open, ready to move ahead an open source version of AI content and start on-boarding contributors soon.