Technology Blogs by SAP
Learn how to extend and personalize SAP applications. Follow the SAP technology blog for insights into SAP BTP, ABAP, SAP Analytics Cloud, SAP HANA, and more.
cancel
Showing results for 
Search instead for 
Did you mean: 
annette_jann
Participant
Chatbots can be found everywhere – often they are used as an alternative channel to answer and prefilter reoccurring user inquiries. Supported by artificial intelligence and natural language processing (NLP) chatbots learn to understand user questions and provide answers in a “natural” way, mimicking human conversations.  

As chatbots mimic human conversations, designing and programming them well is extremely important for the acceptance of the users. Guiding principle is that they should serve a meaningful purpose and not just replicate the underlying application. The benefit of using chatbots within the business context is to serve as the connector or the virtual glue between applications.  

Why A Chatbot? 


In SAP S/4HANA Procurement we heard from our operational purchaser users that they sometimes would love to have a quick way to access procurement information, that might not be directly related with the current task, without having to leave the current context they work in. Just like a quick question to a fellow colleague when sitting in the same room.   

There are already lots of ways on how to provide contextual information in SAP Fiori (e.g. progressive disclosure, autosuggest, situations, ….). Now the question was how might we allow the user punch outs for a quick procurement related info, that is per se not part of the current application without having to leave the current context and needing to refocus again. So, we explored ways on how we could best serve that need.   

Based on our users' feedback, we were looking for a solution that allows to  

  • ask questions, get the right answer, or support narrowing down and maybe suggesting related info (conversational paradigm) 

  • get immediate response (real-time) 

  • understand the semantics right and excuse misspellings (smart) 

  • make use of already known and provided information from previous conversations (context)  

  • grow the capabilities over time depending on the users’ and customers’ needs (pluggable content) 

  • preferably reach out to contacts (MS Teams integration) 


As all of this can be fulfilled by a chatbot, we started working on an Operational Purchaser Chatbot for SAP S/4HANA that provides answers to the most often used queries this role has. The advantage of intelligent chatbots is their capability to understand queries in natural language – unlike a search. This also helps new hires who are not familiar with either procurement processes or a SAP S/4HANA system yet.


SAP Operational Purchaser Chatbot for SAP S/4HANA on the SAP Fiori launchpad


The chatbot is available as a side option in the SAP Fiori launchpad to punch out without having to leave the current working context.  After consulting with our customers, we decided to build this bot as a template, allowing our customers to adapt and extend it to their needs or use it to create their own bots for a dedicated use case. 

Building Bots At SAP 


All digital assistants and chatbots across SAP are summarised under the term “Digital Assistant”.  

To create such a digital assistant “SAP Conversational AI” is a dedicated solution with a building platform, broad coverage of language support and the support for customisable SAP S/4HANA pre-built chatbots. It comes with low-code capabilities like API connections, SAP Single Sign-On, dialog building etc., using state-of-the-art NLP technology to analyse text inputs and to enrich key data. 

The delivered content based on SAP S/4HANA business roles allows customers to adapt delivered templates, extend and even create their own embedded chatbots. 

(For more details check out the article "The future about Digital Assistants and Chatobots in SAP S/4HANA Cloud".)

Designing For Chatbots 


By following the paradigm of learning from the needs of our customers and their users, we first gathered the most frequent queries  for procurement information the users would need as a punch out from the current context. Examples for that could be “Show me suppliers of a PO?”, “Where can I find my contracts?”, “What app to use for posting goods receipt to a purchase order?”: 

Understanding the “what” our users need is sure a good start, but definitely not enough. So we invested a lot in the “how” in the conversational environment as well. 

Although chatbot design has a lot of similarities with design for business applications, there are also specialties to consider.  A conversation between a human and a digital assistant should preferably provide the information the user needs to complete a task easily, fast and on point in quite a small amount of screen real estate. And in addition, it should convey a feeling of understanding and trust – and on top of that aim for an enjoyable experience.  

How to achieve all of the above? The SAP Conversational Experience Design guidelines (SAP CxD) provide lots of information and important design principles to make sure the chatbot at the end satisfies the user’s needs and makes users more productive in their daily tasks. You learn about the importance of providing simple and concise interactions, adapting the conversation to user needs, setting expectations, collaborating and engaging with users and providing accurate and reliable information. In the guidelines you also find various examples that help you in the design process.  

Still to reach consistency and predictability in the interaction and information provided in the bot, and mitigate a few restrictions by the framework, we defined on top clearer UX rules for the card info visualisation and navigation for our chatbot that are following the SAP Fiori guidelines. 

Our Solution 




Our Operational Purchaser Chatbot for SAP S/4HANA (template) is already in its second version and available as of end of February 2022. This example chatbot improves the productivity and collaboration of our Operational Purchaser by allowing to:  

  • View details of purchase orders and purchase requisitions
    You can ask the chatbot to show you detailed information such as statuses or items from purchase orders and purchase requisitions without you having to navigate to the relevant apps. The chatbot is able to show you, for example, all purchase orders from a particular supplier, all overdue purchase orders, and so on.
     

  • Make calls using Microsoft Teams
    You can use the chatbot as a convenient way to initiate an ad-hoc collaboration via Microsoft Teams. So you can, for example, reach out to a supplier via the contact data associated with them. 



  • Schedule meetings using Microsoft Teams
    You can ask the chatbot to schedule meetings in Microsoft Teams. The chatbot is able to retrieve all necessary data from Microsoft Azure AD on your behalf and create calendar events using Microsoft Graph.

  • Navigate to Fiori apps
    You can drill down for details of purchase orders and purchase requisitions by clicking them in the results displayed in the chatbot. This will open the item in the relevant SAP Fiori app. For other related business objects the chatbot can help guide you to the relevant apps. The navigation functionality is able to understand natural language with no need for exact app names, but simply by describing what you want to do, for example asking where you can view all purchase orders. 


The Operational Purchaser Chatbot for SAP S/4HANA (template) provides our customers a great example on what's possible when creating own chatbots for their SAP S/4HANA system using the SAP Conversational AI service, allowing them to re-use, adapt and create own bots based on it. 

Thank You Team 


Our small cross SAP S/4HANA Procurement project team, consisting of development, architecture, PM and UX from the incubation team, as well as user assistance, worked in in a truly agile way.  We jointly explored the capabilities of SAP Conversational AI, technical possibilities and constraints, availability of data in the system, learning mechanisms, existing SAP Conversational Experience Design and user assistance guidelines, …   

On demand we connected with the experts from business applications, SAP Conversational AI and the central SAP UX – CxD team responsible for the SAP Conversational Experience Design Guidelines, who provided support where needed and appreciated our feedback and ideas to feed them back into the framework and guidelines.  

We would like to thank all involved colleagues for the great cross team collaboration to reach meaningful and enjoyable solutions for our users – exemplifying once again SAP S/4HANA Procurement loves UX. 

Co-Authors: nikeklaubert and annette.jann