AI at SAP: The Path to the Intelligent Enterprise

Is unlocking your smartphone with facial recognition the ultimate convenience or simply a curse? It’s a question anyone who has set up a new smartphone is likely to have asked themselves. Artificial intelligence (AI) not only makes life easier for consumers, it has also become a mainstay of many enterprise applications.

For SAP, AI plays a key role in empowering its customers to become intelligent enterprises. The main motivation of implementing artificial intelligence is to support people – freeing them from mundane, repetitive tasks and thus giving them more time to employ human skills, such as creativity and empathy.

But whether in business or the personal realm, there are ethical issues that need to be addressed. SAP first defined guidelines for using artificial intelligence in 2018 to ensure that ethical issues and conflicts are considered in the development of new services and use cases from the very beginning. And the SAP Global AI Ethics Policy, which applies to all employees, came into force in January 2022.

In addition, a committee of experts and an external advisory council also review questionable cases on a regular basis. Facial recognition was already an item for discussion here. For example, an AI service within an industry solution was intended to identify people and warn them if they entered a danger zone in production – which is a good thing, in principle. However, the same technology could be used by an unjust regime to identify and monitor its citizens, for example, and this usage contradicts the principles of ethics policies at SAP. This technology was never implemented.

Key Technology for the Intelligent Enterprise

Analysts estimate that the global market for artificial intelligence in software grew by 20%-22% from 2021 to 2022. Forrester predicts that revenue in this software market will grow by 39% from 2021 to 2025. Clearly, artificial intelligence is a true megatrend. SAP recognized this potential several years ago and intends to become a leading provider of enterprise AI for key technologies.

The company invested in its artificial intelligence business unit, which is led by Dr. Feiyu Xu, a researcher and experienced AI expert. Potential multipliers here include the company’s strong market position, access to the business data needed to train the AI, and profound knowledge of business processes in all industries. SAP is focusing on embedded AI: services that are part of an industry solution and work in the background to support users. SAP also offers individual artificial intelligence microservices – SAP AI Business Services – as well as AI infrastructure and tools – SAP AI Core and SAP AI Launchpad – with which SAP’s business units, partners, and customers can implement their own artificial intelligence projects.

“Customers expect the solutions they buy from us to be intelligent by design,” says Jana Wuerth, product management expert in the AI unit. “In the best case, end users don’t even notice that the solution contains AI components, but only that it works well and makes their work easier.” So far, several hundred use cases have been developed in many different business areas. Customers include Hewlett Packard Enterprise, Dulux, Mitsui, and Döhler, as well as many midmarket companies. Customers know that these services work best when experts use real customer data to train and test them. That’s why more and more customers are making the data from their systems available for development – under strict data privacy criteria. “We are happy that so many customers are sharing their data with us, because it helps us make their applications even more efficient,” says Dr. Xu, SVP and global head of Artificial Intelligence at SAP.

AI Created on an Assembly Line

The key to making this happen is the close collaboration between lines of business and AI technology teams. The SAP collaboration model “AI Factory” equips applications with AI like on an assembly line. With this approach, SAP can provide focused solutions, delivering ready-to-run artificial intelligence and other services as a one-stop shop that customers can use simply, efficiently, and uniformly.

The team is currently working with its line of business colleagues on several use cases for order automation. For example, customers who have not yet automated their orders, but instead receive them as PDF files, used to have to enter the data in their system manually. An AI service can now extract the necessary information from the PDF and create a PO automatically. A planned future addition is a service that supplements missing information autonomously by reading it from the order history. This could involve the order type, plant, or route, for example. If information was missing, employees used to have to inquire with their contact person; this step has been largely eliminated. Depending on the quality of the input data, the accuracy of some of these services is nearly 90%. As a result, manual processing or review is only needed for a small number of orders.

Another AI service from SAP makes personalized recommendations possible, as consumers are familiar with from online shopping. They are also used for SAP’s internal course offerings in its HR development software SAP SuccessFactors solutions, with employees receiving relevant training offers based on their business role and the courses they have already completed.

Intelligence in SAP Process Automation combines features of the SAP Workflow Management service and SAP Intelligent Robotic Process Automation with a powerful, intuitive, no-code development environment. This allows users who have little experience in software development to design, change, and improve their own workflows and business applications. In addition, embedded AI supports the business processes, making them smarter and more efficient.

One solution for the circular economy is SAP Returnable Packaging Management. When it comes to returnable packaging, such as shipping pallets, heterogenous data systems have made it difficult for the different players in the supply chain to collaborate. The AI solution, which builds on SAP Business Technology Platform, now helps integrate the different back-end systems.

Myriad Possibilities

The healthcare line of business is currently exploring an AI service for hospital billing. As background, the German legislature made the subsequent correction of bills more difficult at the start of 2022, which makes it essential for healthcare providers to issue correct bills in the first place. If an employee in the billing department enters “appendectomy,” for example, the system automatically proposes line items that are typically involved in this type of treatment, such as specific medications and services. This can not only simplify and speed up the billing process, but also can prevent the creation of incomplete or erroneous bills.

Many more use cases for artificial intelligence are imaginable in the future. Integrated AI in an industry solution could predict interruptions in the supply chain, for example, because it has been trained with real data and can identify patterns that resulted in bottlenecks in the past. Companies could also use AI in the future to avoid procuring products produced with child labor because the procurement solution has been trained to identify such risks in the supply chain at an early juncture and search for alternatives.