With SAP intelligent RPA, the vendor aims to boost S/4HANA's automation capabilities with prebuilt bots.
Robotic process automation -- or RPA -- has been one of the fastest-growing segments in enterprise software because it facilitates the rapid automation of many tasks by simulating the way a human interacts with apps.
SAP began working on Intelligent RPA technology in early 2018. It acquired Contextor in November 2018 to augment these early efforts with connectors to non-SAP tools, web apps and Microsoft Office. The first version commercial version of this new product was launched in May 2019.
"Due to this and additional investments, SAP managed to become a relevant global RPA player in just two years," said Björn Bartheidel, head of digital manufacturing and IIoT at managed cloud services provider Syntax.
SAP intelligent RPA, sometimes referred to as iRPA, is meant to provide better integration into the SAP application stack than other platforms and easier implementation when companies have adopted SAP-prescribed best practices for enterprise processes such as procure-to-pay or quote-to-cash. Organizations can also use intelligent RPA technology as a complement to other RPA platforms that may provide better automation across non-SAP software applications.
"Many organizations adopting RPA adopt more than one platform," said Maureen Fleming, vice president of intelligent process automation for IDC. "Because many companies adopting RPA start by automating financial processes, it wouldn't be at all surprising to see SAP intelligent RPA competing with other RPA platform vendors as well as see enterprises running both, depending on the use case."
The technology has a number of benefits and use cases that companies are exploring.
The Asia-Pacific office of Rehau, a polymer manufacturer, distributor and designer, recently adopted intelligent RPA.
Intelligent RPA provides deep integration into the organization's existing SAP backbone, said Chengbo Yu, CIO for the Asia-Pacific region of Rehau. The platform is transparent to business users who are already familiar with SAP. As a result, developing and testing new automations takes less work and requires less of IT's involvement. RPA projects can move into production in under two weeks, which enables Yu's team to adopt an agile, iterative approach in improving both the automations and the processes they automate.
As a key to this success, keeping each bot relatively simple is important, Yu said. A quick win provides immediate business benefit and improves buy-in from the business. The focused, simple approach also makes the ROI more transparent.
Automating best practices
SAP's current strategy lies in positioning intelligent RPA as a complement to other RPA products and the company is actively partnering with RPA leaders such as UiPath, said Oliver Betz, senior vice president and global head of product management for S/4HANA at SAP. The real strength is pre-building an RPA component into S/4HANA applications, he said. This can help automate the process of creating RPA automations -- that is, bots -- with less work. SAP launched a program in 2018 to support 10 pre-built automations and now offers about 100.
For example, if a customer is using a standard invoice-to-pay process with standard SAP content, that customer can use an RPA bot out of the box. One of the benefits of this is that these standard processes enforce the best practices observed by SAP from work with the industry in various domains including invoice-to-pay, procurement, sales, professional services and manufacturing.
These are all processes implemented across SAP applications. "We would never do a bot that does a process step in Microsoft Dynamics or Salesforce since we don't have expertise there," Betz said. "Our core proposition is in automating S/4HANA process steps."
The design of the SAP Cloud Platform organizes discrete capabilities so that the individual technologies can work together seamlessly to solve a larger automation problem.
"That creates strength for iRPA because customers are able to do more with the tooling," Fleming said. For example, intelligent document processing is a layer that most of the RPA vendors are beginning to offer as they extend capabilities to support AI use cases. SAP has a document processing service that works with iRPA and is callable by a workflow, a form or by an integration component.
Embedding intelligence into workflows enables them to get better over time, Betz said. SAP has already developed modules for procurement, sales, service and inventory management that can spot and automatically fix inconsistencies in data using historical data stored across systems. For example, a company name might be corrected to "Johnson & Johnson" from "Johnson and Johnson" to ensure consistency when a user realized that all previous purchases were bought from the first designation. The module can make corrections where a high level of confidence exists, or it can make an update and flag this for a double-check by a human.
The development tools in intelligent RPA are intended to be well-integrated with SAP applications.
Individual components of the UI are addressable from the studio library and can be dragged into an automation, said Fleming. This makes it easier to use for inexperienced end users trying to automate. It also improves resilience, creating a separation of UI from individual automation projects.
If the target application UI changes, the RPA team can make changes in intelligent RPA in one place without having to manually change each automation that needs to interact with that application.
While reusability is not entirely unique to SAP, reusability is a strength of SAP's bot development tools, Fleming said.
SAP intelligent RPA does pose some challenges, however.
Enterprises should keep an eye on the cost to run automations, Bartheidel said. SAP's licensing model is transaction-based.
"Transaction-based licensing requires internal governance and controls to avoid unexpected and unplanned costs," he said. "On the other hand, it allows a small-budget entry into the world of digital workforce."
Another limitation is that it only works via the cloud. Only being able to access the tools via a browser could limit the potential usage, Bartheidel said. This could be a bigger problem for areas with slow or even no internet connection.
Also, intelligent RPA does not have as many prepackaged activities as competitive RPA offerings, he said.
"They are rapidly working to package, but so is everyone else," Fleming said.