2 Top Limitations of Implementing RPA in the Enterprise

In this article, Jakob Freund, co-founder and CEO, Camunda, shares an adverse perspective on the limitations of implementing RPA in the enterprise. He also discusses the RPA landscape more broadly, including how it compares to other BPM solutions, low code development, AI, and other workflow automation technologies.

In 2021, the robotic process automation (RPA) sector has heated up, with M&A and funding activityOpens a new window proving the market’s demand for automation. With the pandemic driving a whole new sense of urgency around enterprise digital transformation, many teams may be tempted to automate their workflows with RPA. While these efforts are well-intentioned, enterprises may find themselves limited by the capabilities of these tools in the long run.

When intertwined with legacy systems and modern, cloud-based applications, an RPA implementation can only go so far. Here’s why that’s the case and how to use end-to-end process automation and orchestration to fuel enterprise modernization efforts.

1. Balancing Short Term Needs vs. Long-term Priorities

Some things about RPA may seem too good to be true, including opportunities to increase productivity, minimize human error, and streamline complex workflows. The issue is that many teams use RPA as a way to get around legacy technology (think decades-old COBOL systems that are still in production, for example). That’s often because ripping and replacing legacy systems is disruptive. This practice can cost thousands (if not millions) of dollars and take years to complete. 

While it’s tempting to want to rewrite every part of the technology stack as a modern, microservices-based application, doing this all at once can grind an enterprise’s critical processes to a halt. To get around this issue, enterprises take to RPA systems to automate small actions, such as re-typing results from a green screen into a web interface or getting data from scanned paper documents into a CRM. 

It’s important to understand that this is a short-term fix. An RPA bot programmed on top of a legacy system lacks the underlying API connectivity for deeper, more complex automation. This bot can simply automate and control the existing user interface (such as the green screen COBOL application mentioned above). Unfortunately, 91% of teams using RPA in 2020Opens a new window reported experiencing problems with implementation, such as managing security, maintenance, and lack of control.

See More: 4 Ways Pairing Lean Manufacturing with Automation Can Mitigate Labor Shortages

2. A Poor Developer Experience

While a low-code implementation process may seem like a benefit at the outset of an RPA deployment, developers are often called upon to help when things go wrong. As event-driven processes get more complicated, RPA deployments can become bottlenecks or even fail altogether as enterprises wrestle with legacy systems. That’s where the developer’s job of untangling these low-code RPA deployments begins.

Because RPA bot architecture is often disjointed and task-based, the development team may not have visibility into why, where, and how a more complex process might break down. Many enterprises use both legacy and SaaS applications within the same departments. These systems don’t play nicely with one another or seamlessly share data where it’s needed. 

Ideally, automation tools need to be built for developers, mixing low-code shortcuts with the capability to drill down to make custom code-level changes where needed. A developer-friendly approach should start with an open and flexible architecture rather than more rigid RPA workarounds. 

A Stepping Stone to Modernization

Many enterprises may find themselves already in too deep with RPA, which isn’t necessarily a problem in and of itself. The smartest way to contend with these deployments is to separate the orchestration layer from the RPA bot layer. Think of an orchestration layer as the software that “drives” one process into another. Since processes rarely consist of only one step, orchestration can ensure that the chain of events remains unbroken and give IT teams more visibility into where processes break down.

To take that modernization a step further, development teams can gradually replace RPA bots one by one with applications built on a more modern, microservices-based architecture. This gradual digital transformation may make sense for most enterprises that don’t want to rip and replace legacy systems all at once. It’s far less disruptive to the business and allows developers to re-architect and modernize business applications in priority order based on what is most urgent for the customer experience.

Once applications are rewritten as microservices, it’s much easier to make changes, connect applications via APIs, transfer data, and more. Orchestration software works much more seamlessly from end to end when this API connectivity is in place. Not to mention, the IT team can get the visibility they need to detect process breakdowns and continuously improve the efficiency and effectiveness of enterprise processes. 

See More: Low-Code Automation: How Businesses Can Avoid the Pitfalls

The Need for End-to-End Process Automation and Orchestration

Often, RPA deployments seem too good to be true…because they are. Enterprises have ambitious modernization goals, paired with a mix of complex legacy systems and modern, cloud-based applications. Combine that with the fact that IT leaders and business stakeholders often can’t see eye to eye on how to prioritize. Luckily, using open standards like business process model and notation (BPMN), these teams can collaborate using easy-to-understand, flowchart models to annotate their processes. Within these models, they can determine where people and technology interact and identify potential issues or bottlenecks. From there, they can focus on which automation efforts to take on first, based on what matters most to the customer experience. It’s not just about quick fixes. It’s about automating and orchestrating business processes from end to end.

Did you find this article helpful? Tell us what you think on LinkedInOpens a new window , TwitterOpens a new window , or FacebookOpens a new window . We’d be thrilled to hear from you.