SAP has developed and refined its approach to robotic process automation (RPA) for some years now, stretching back to a time (half a decade ago perhaps) before this strain of technology became one of the new “darlings” of the tech landscape. In its evolution, SAP has been quite specific about putting an “i” (for intelligent) in front of its offering of RPA.
We know that RPA at its most basic level has been described as “just screen scraping.” One example might be tracking a user’s actions on the screen to see what fields in a forms-based application are interacted with and then programming a bot to emulate those actions when the same task is required.
If this is RPA 1.0 and the notion of SAP to progress RPA into functional workplace environments was RPA 2.0, then we can now look to a more robust approach for enterprise-grade process automation as we now move to RPA 3.0 in 2020 and beyond. So, what does this encompass?
What Is Enterprise-Grade RPA?
For RPA to be considered enterprise grade—battle-ready and robust enough to deliver in mission-critical commercial environments—it needs to emerge out of its adolescence.
This is RPA that runs on data models that have been tested. In the simplest terms, a data model describes the relationships between all the entities in an organization’s data domain (the database of choice, most likely SAP in our case, plus other cloud-based data stores) and what those entities mean to each other (e.g., how they are related, what their core attributes are, and how they interact). Although abstract in nature, we need to have a very real view of our data model before we trust it to fully automate.
When Can We Trust the Bots?
This is RPA with governance. From the data model, we need to know what data compliance and regulatory controls our data must to adhere to. Once we know our data is compliant, then we can be more relaxed about handing it over to the bots.
This is RPA with data and process provenance. This is the people factor—if we start trusting citizen data scientists to create functions in the workplace, then we need to know who created what, when, and where. Policy controls should be in place to ensure unqualified people aren’t logging in and creating go-to-market strategies that manifest themselves as bots.
SAP RPA for Automated Business
SAP itself reflects a lot of those messages, but initially only at a higher level. The company’s core RPA messages suggest that SAP Intelligent RPA can automate repetitive manual processes by creating, scheduling, managing, and monitoring intelligent bots. ASUG readers will note that SAP listed management toward the end of that list.
This RPA pizzazz has arguably come about because software vendors like SAP are still keen to sell the sizzle and promise of bot technology. “RPA and the broader theme of automation is a hot topic for enterprises, with strong competition between vendors to help meet those enterprise needs,” said Tom Pringle, VP of technology research at G2. “As the potential use cases for RPA continue to expand, so must the enterprise-friendly features and functionality that support them. Extending RPA by integrating AI-powered capabilities to enable more complex automation, backed by native governance features that helps address misuse concerns are examples of logical, enterprise-friendly additions that will help RPA use become more pervasive.”
SAP promises to “Develop bots for error-free, scalable tasks and help you focus on high-value processes. Cut across organizational barriers while driving optimal process models for maximum efficiency. Respond to customer needs proactively and augment resources.” There is an SAP Intelligent RPA administration guide with a lot of the identity, trust, and security control technologies needed, but you have to look for it.
Key Technology Drivers for Good RPA
If there are three key drivers for good RPA, it might be the following factors:
- The power of three in connected orchestration: a coming together of all three “work agents” of modern work—humans, bots, and artificial intelligence (AI)—to make sure that all three corners of this triangle work in unison and in harmony toward collective and mutually beneficial goals.
- Human-in-the-loop: When software bots create errors or exceptions, humans are involved to make corrections immediately. This means the use of a business interface for managing bots. Tasks will include monitoring, scheduling, and reporting (available on web and mobile devices) to deliver analytics and reporting ranging from impact analysis to compliance.
- AI in the mix: Good RPA contains plug-ins to sophisticated AI engines from partners like Google or IBM.
RPA Escalation and Growth
According to analyst house Forrester, the RPA market is expected to reach $12 billion by 2023, but many organizations are still in the early stages of their RPA journey. To date, we can see that many customers have experimented with RPA by implementing bots from multiple vendors, resulting in a fragmented technology landscape. This is not RPA 3.0 at an enterprise-grade level.
Only when we build enterprise-grade RPA 3.0 can we achieve trust—without trust, the bots just won’t make the grade.
If we are going to bring RPA technology forward into the workplace, we will need to have the sometimes uncomfortable discussion as to whether or not it is meeting the enterprise grade. Going forward, we may see some differentiation between “user bots” or “hobbyist bots” made by users who are experimenting, and those that have been officially ratified as fit for purpose in the business. Let’s call those “big bots,” just to see if it sticks.
Learn more about the pros and cons of RPA in the on-demand ASUG webcast, “Automating Financial Processes Through Robotic Process Automation (RPA).” Register for the ASUGFORWARD Virtual Experience taking place June 22 – 25, 2020 to learn, exchange ideas, and gain practical information you can put into practice immediately.