In early December we had a Hackathon with SAP partners that was focused on using Intelligent RPA with Ariba. You can view the summary of that Hackathon in this blogpost from the Co-Innovation Lab. This Hackathon challenged partners to design and develop a use case for automation with Ariba. Over the next few weeks I will be sharing use cases in a series of use case spotlights. Last week I shared the use case from PWC which automated the contract workspace, you can view that post here.
For the second post in the series we will be looking at the use case from the Capgemini team. The process the team focused on was the requisition creation process in Ariba. By implementing SAP Intelligent RPA the team built a bot that enables users to order supplies through email or with a chatbot. By enabling these two alternatives for ordering the team can improve the user experience for the employee while lowering costs by favoring catalog items. The Capgemini team estimates this bot saving 40% on ordering. The team impressed the judges with their chatbot and requisition automation and was third place in the Hackathon.
This bot was designed to allow users to interact with it in two ways. A user can chat with a chatbot to place their order, or send an email to the bot's inbox with the order information. By using the SAP Conversational AI as a chatbot the user experience becomes a conversation. If a user needs to submit a larger order, or doesn't want to use the chatbot they can instead utilize the email input and submit their order request to the bot via email.
Once the user has submitted a request via email or with the chatbot, the processing bot is triggered. This processing bot is triggered with the API trigger. The API trigger is one of three ways to trigger a Intelligent RPA process and is well suited for a use case like this where the bot is utilized on an as needed basis instead of on a schedule. When the bot is triggered it logs into Ariba and creates the requisition using the details from the user. After the bot creates the requisition it captures the requisition ID and other required information and sends it to the user via email. You can see the solution architecture from Capgemini below:
With this automation the end user is able to place an order quickly in a very user friendly format, and continue with their day while the bot handles all the required steps to submit the order. Human and bot collaboration is a great way to utilize a tool like RPA, and this example shows how a bot can be implemented to improve the user experience while simultaneously improving the process.
A topic that is becoming more common with RPA is the concept of hyperautomation. With hyperautomation the goal is to automate an end to end process that is adaptable by leveraging new tools like machine learning (ML), robotic process automation(RPA), and natural language processing (NLP). With the inclusion of the Conversational AI which uses NLP this use case is an example of hyperautomation. The chatbot being used as the source for ordering information allows this bot to be improved without impacting the underlying Ariba automation. For example, permissions could be added to the chatbot which limit what certain users can buy, and the API call and Ariba workflow wouldn't require any updates. Frameworks that allow improvements on a modular basis reduce the cost of changes, and can provide opportunities to reuse existing workflows. With automation being a tool to reduce repeated tasks, there is no reason to repeat development for a workflow. Recently sebastian.schroetel, Head of SAP Intelligent RPA had a discussion about hyperautomation and the hype around it amidst a global pandemic. You can read that discussion in this post.
By automating this process the bot has been designed to prioritize catalog items, so when a user submits an order it will always search for that item in the catalog before ordering it somewhere else. This improvement is part of the savings that this bot can provide, in total the team estimates roughly 40% cost savings by implementing this bot.