There’s a burning question around SAP’s next move in automation. The short answer is that it will be highly automated.

As the trend toward software automation takes shape in SAP’s development road map, SAP will likely embrace the opportunities that automation gives us to connect users across different applications without the burden of APIs and other “connective tissue” forms of software engineering. Automation technologies today exist to learn from user (and machine) behavior, which means they can form new logical connections without the need for complex scripting and APIs.

As described in “SAP Explains its Foray into Robotic Process Automation (RPA),” the bulk of technologies in this space gravitate toward classifying, recording, and monitoring data as it travels through workflows inside applications and database operations that form business processes. Once recorded and monitored, business process automation can occur. A classic RPA model runs on a record, build, run, and monitor cycle. But where does it go from there? And where will SAP take it? 

Making Automation Smarter

SAP has sought to perform enterprise-level RPA. It has always referred to automation’s presence inside SAP HANA, SAP Leonardo, and across SAP Cloud Platform as part of an intelligent system. For this reason, SAP has always called it Intelligent RPA.

Intelligent automation, as SAP sees it, is a means of replicating human actions that can be accurately codified by software and executed with accuracy. Yet that automation still needs to know that there are known unknowns. It should be programmed with enough know-how to realize what it needs to learn about wider and extended business processes, in order to get smarter.

Two Levels of Bots

SAP’s Intelligent RPA will often manifest itself as a software bot. But this is a bot at two levels. Some bot intelligence might be presented in a graphical user interface (GUI) that a user interacts with, which is often called “attended RPA.” Equally, some automation might not need to touch the user until later, so the bot’s intercommunication happens at a machine-to-machine (more likely application-to-application or database-to-database) level. The resulting efficiencies will ultimately be presented further down the line in a dashboard, as a data query, or in an application function that is smarter as a result of the Intelligent RPA DNA feeding it.

We’re getting used to conversational bots on the websites we visit everyday. Love them or hate them, they often start as fully automated services that perform a “human handoff” to a support engineer once the bot runs out of answers. These bots must get more intuitive on the road ahead if we expect wider user acceptance. Equally, it’s comparatively early days for machine-to-machine bots. Software engineers will be looking to bring greater levels of automation efficiency forward with back-office bots on the immediate road ahead. If these bots aren’t on the average data engineer and developer’s architecture plan today, then they will be tomorrow.

This latter type of action is often called “unattended RPA,” especially when it is related to autonomous robots deployed in server farms to perform back-office processes that are automated from end-to-end.

So, if we know that SAP already has an established foundation in automation across its base of platforms, what aspects of feature development and innovation can we expect the organization to push toward next?

Weeding out the Intelligent RPA

The next era of automation features a certain amount of weeding out of existing business processes before we can plant the new ones. We’ve heard this kind of thing before—for example, cloud computing is for everyone, but not for everything. Some applications and data should not be virtualized over cloud platforms for regulatory reasons, latency, or privacy concerns. Places, areas, and instances where it’s not a good idea to automate include:

  • Design planning (machines don’t have human creativity)
  • When subjective software testing is needed (machines can’t express “look and feel” very well)
  • Complex interwoven functions (cost may outweigh investment if existing systems work OK)
  • Where the automation function only runs once (in areas like app testing, we need to look for often repeated tasks to justify efficient use of automation)
  • Where applications are still unstable and in more embryonic stages of development.

Similarly, the article mentioned above states that Intelligent RPA can work in every department, but not for every business process. Some elements of business will, for now, require highly personal, highly customized, and highly human functions with levels of empathy and nuances of interpretation and intuition that we cannot expect computers to emulate.

Playing by the Machine Rules

As we know, machine learning is the practice of teaching a computer how to spot patterns and make connections by showing it a massive volume of data. This allows Intelligent RPA to tackle repetitive, rule-based (often monotonous) tasks. But knowing where to dig and plant Intelligent RPA is an important initial step.

SAP’s vision of RPA techniques for the intelligent enterprise with SAP S/4HANA as the digital core includes SAP engines. These are SAP software components that provide automation relying on highly specific process knowledge. But it’s important to remember that engines have a fixed level of logic and a limited configuration possibility. This means, realistically, that we can only code a specific number of “business rules” into any one engine, meaning that we can never cover all facets of the business processes.

Where code-based software engines can’t give us all the answers (they are estimated to have the potential to facilitate automation in up to 60% of all cases), we need to turn to the wider worlds of machine learning and RPA that we’re discussing here. SAP points out that its cloud services operations management and maintenance can be fully automated, but that a stable solutions portfolio is also an integrated solutions portfolio—in SAP’s mind, ideally an integrated SAP solutions portfolio.

Automating Through Cross-Platform Complexity

Another key area of development is cross-platform Intelligent RPA integration. Again, there’s quite a lot of digging and weeding needed here to make sure we’re set for productive business. SAP bought Contextor at the end of 2018. The company’s technology is designed to simplify user interface interactions across disparate systems. SAP will use it to simplify Intelligent RPA user interface interactions across SAP and non-SAP applications.

It’s clearly early days with Contextor inside SAP. But when SAP does bring it into the fold in more concrete terms (presumably as part of SAP Leonardo), it will no doubt be focusing on the “mechanics” of making cross-platform Intelligent RPA happen.

These mechanical steps will typically include:

  • The classification and measurement of business process tasks
  • The ordering, sequencing, and planning of tasks
  • The ongoing contextual data collection of tasks, especially where cross-platform non-SAP RPA needs to be channeled into an SAP application
  • The need for compliance checks and validations across tasks

Automation in Practice

The Contextor Interactive Assistant product aims to perform a version of the above list. It also allows the fluent navigation between applications using Intelligent RPA technology. Once again, it’s all about rule capture. Where we can capture digital business and ascribe accurate numerical values to it, then we can apply automation to its maximum effect.

As something of a final caveat, it may be worth noting that it is potentially possible to apply too much Intelligent RPA, however intelligent it is. CIODive recently quoted Eric Stine, chief innovation officer at SAP North America who said, “Companies looking to succeed in automation require a combination of speed and humanity to guide their actions.”

Stine’s comments resonate with the “intelligence still needs humanity” message we touched on earlier. People don’t want to feel like robots are replacing their jobs overnight. They need to be shown (at their own pace) how Intelligent RPA can make their lives better and allow them to do higher-level, more rewarding workplace tasks.

The ERP Intelligent RPA Road Ahead

With all these technologies in a state of flux inside SAP central headquarters, one can only wonder what kind of message we’ll hear at next year’s SAPPHIRE NOW and ASUG Annual Conference.

It’s not unreasonable to imagine that SAP will want to use SAP Cloud Platform to bring together ERP from its own established base of innovation and experience, plus cross-platform Intelligent with Contextor, plus experience management from the recently acquired Qualtrics, to form one new total offering for us.

Want to learn more about the pros and cons of RPA? Register for ASUG’s on-demand webcast, “Automating Financial Processes through Robotics Process Automation (RPA).