Skip to Content
Technical Articles

Combining CAI and Intelligent RPA Bots to Solve Your Most Complex Problems

INTRODUCTION

SAP’s Conversational Artificial intelligence (CAI) platform offers developers powerful tools to create innovative chatbots, however they become truly game-changing when they are paired with an Intelligent Robotic Process Automation(Intelligent RPA) bot. The two bots come together to create a “super bot” of sorts that can solve problems that would stump an individual bot. The CAI bot acts as a front end, interacting intelligently with the user, gathering data based on conversational parameters. The Intelligent RPA bot on the other hand acts as the back-end workhorse, executing manual actions at lighting speed whenever a user completes a conversation. Theses back-end actions are often performed by API requests however there are numerous scenarios in which Intelligent RPA is the most cost effective, preferred or just straight up the only option available for development team. To get an idea of how CAI + Intelligent RPA can bring huge composite value let’s take a look at a user deprovisioning chatbot for an ancient legacy system:

USE CASE

  • Users need to deprovision themselves or another user within the system.
  • Admin team reports abundant errors in deprovision requests when done manually
  • The deprovisioning system is a custom-built legacy application with NO API SUPPORT

CAI BOT

Using SAP’s CAI platform, we can build a chatbot which is able to detect the intent, mood and desires of the user. It uses this information to gather up the necessary information for a deprovision request: user id, last day, archive location, etc.:

 

  • For example, we can simply configure the bot, using the CAI platform, to ask the user for a #datetime entity, like so:

.

  • Therefore, if #datetime is missing from the conservation memory the bot will prompt the user with the following:

  • Once the bot has completed its conversation with the user it sends the gathered information from its memory in an API call to the Intelligent RPA Cloud Factory where the Intelligent RPA bot is then triggered. This webhook call is setup like so on the CAI platform:

 

INTELLIGENT RPA BOT

  • The Intelligent RPA bot is used to capture the various screens of the deprovision system and is developed locally on a developer’s PC:

 

  • In order to enable API data input for the Intelligent RPA bot you need to navigate to the Memory Context Menu, create a new folder called “IO”, a create a sub-folder to that called “Input”, and then declare the variables you want to be available to the API there:

 

  • Once the bot has been created it gets imported to the Intelligent RPA Cloud Factory where it is then deployed:

 

  • Once deployed, a trigger of type API is added:

 

  • This is the information that is generated by the system once the API trigger is created. We now copy & paste this URL and payload into the CAI platform webhook call[shown in an earlier step] and we are good to go!:

CONCLUSION

Combining all the above components will create a seamless user experience, where the CAI bot holds the users hand through the deprovision request, and then the Intelligent RPA bot takes over to do the back-end work of submitting the request manually because of the legacy systems LACK OF API SUPPORT. This creative solution provided tangible benefits to the company in remarkable short time due to the simple development processes of CAI and Intelligent RPA.

Conversational AI and Robotic Process Automation are a match made in heaven. They complement each other’s strengths and they cover each other’s weaknesses. The problems they can solve together and the applications they cover for are only limited by your imagination!

REFERENCES

https://help.sap.com/viewer/c836fab4182e45548b6c6c6d0d0a9146/Cloud/en-US/f2e3d2d3e84d42eabe14913f7ee9269c.html

https://help.sap.com/viewer/c836fab4182e45548b6c6c6d0d0a9146/Cloud/en-US/347ba172a8e64f1c8544f46026446a86.html

https://cai.tools.sap/docs/concepts/code-and-webhook

 

6 Comments
You must be Logged on to comment or reply to a post.