As of recent, MuleSoft is expanding its capabilities offered across integration, API management, and automation, making it even easier for organizations to execute their digital transformation agendas and build composable enterprises.
A recent Salesforce study finds surging demand for automation, and even though RPA has gained enormous market traction over the past few years, many organizations are just now starting to learn and evaluate the technology. We’re here to demystify RPA, broaden the aperture of RPA’s business value, and share common challenges that organizations face when trying to scale this technology.
RPA 101: Challenges and opportunities of robotic process automation
At its core, RPA is a software solution that is used to create RPA bots. RPA bots are scripted or “trained” to imitate the manual actions of a human across a business process and by doing so, automate the process.
RPA has been a top priority for an organization’s digital transformation journey over the past few years and may be even more important today given the organizational and operational shifts that COVID-19 has thrust upon organizations according to a recent Gartner study. And why not, considering that an organization can significantly increase its productivity and efficiency by running business processes 24/7, 365, all without the need for human workers.
That said, not all manual business processes are good candidates for RPA. As a starting point, a suitable process for RPA should be:
- Repetitive: The value of RPA is driven by the automation of manual, time-consuming activities. For example, an organization will achieve more value by automating an hour-long daily process versus an eight-hour process that is executed quarterly.
- Rules-based: Human interpretation should be minimal throughout the process, as standalone RPA lacks cognitive capability to make judgment-based decisions. That said, a business process is never 100% consistent. Rules-based objection-handling can be included for known process deviations from the primary “happy path” of the business process.
- Standardized: The process should not frequently change. Similar to the previous point, processes should follow a predefined path or decision tree.
- Digital: Inputs and data across the process are stored and/or can be accessed digitally. Best case scenarios involve enterprise applications like ERP and CRM versus local drives, locally maintained spreadsheets, etc.
Given the above points, back office functions like Finance, HR, and IT are often ripe with RPA opportunities. But even if a process checks the above boxes, RPA may not be the silver bullet.
As a simple example, suppose an analyst pulls daily sales data but each time must spend 30 minutes rearranging data columns to execute further analysis. While RPA is certainly one solution that could eliminate the analyst’s time spent on rearranging data columns, it is probably not the best solution in this scenario. It is likely that the data warehouse has features or business intelligence solutions that can be leveraged to extract data in the structure required by the analyst. As with all process evaluations, consider other existing alternatives before moving forward with RPA as the solution.
The [ro]bots are coming here
Forrester estimates the current RPA software market at $2.4B, over an 800% increase since 2016. Simply put, RPA has exploded over the past few years.
Even with this meteoric rise though, very few organizations have achieved sustainable enterprise scale. While there is not one exact reason to point out, we can call out a few specific challenges.
Taking the “faster” route
To achieve faster speed to deployment, building RPA bots initially largely relied on graphical user interfaces (GUI), which are the type of interfaces that we interact with via mouse/trackpad for example.
The challenge often faced with this approach is that the layout of an application’s GUI is not always fixed. Consider how often companies refresh their UI designs or update their software which can also impact GUI layout. As an example, suppose a bot was originally trained to click on a button located in a specific location of an application’s GUI.
However, that same button slightly moves position after a software update. In this case, the bot would likely fail and need to be retrained. Therefore, deploying RPA bots that leverage GUIs may be faster, but the risk of greater ongoing maintenance increases, which could eat away at the value gained from automating the process to begin with.
Another approach that provides greater, sustainable value relies on application programming interfaces (APIs). With APIs, bots can execute commands behind the screens, eliminating wait time for screen loads and being immune to changes of an application’s GUI. Additionally, deploying RPA in this approach provides greater flexibility, adaptability, and security that is inherent in composable API-led architecture. However, the downside is slower time to deployment if APIs do not exist for target applications.
From simple tasks to complex workflow automation
Organizations often find quick success automating simple tasks and other low hanging fruit such as manual data entry. As an organization’s RPA capabilities continue to mature, the natural next step in the RPA journey is to automate end-to-end processes and workflows. However, going from single task automation to complex, multi-step processes is a giant leap.
A multi-step business process exponentially increases the number of integrations, decision points, security controls, users, etc. that are necessary for automation. Additionally, is it likely that more than one bot is needed to automate an end-to-end business process. Even if only one bot fails, the entire process likely fails as the subsequent bot(s) in the process may not be triggered. Accounting for all these points, the number of potential failure points is much higher with end-to-end process automation.
That said, one way to mitigate failures associated with complex business processes is to invest heavily in the upfront RPA lifecycle steps of process mining and design. During this stage, Business and IT work together to greatly detail the primary process pathway and decision points, while also identifying all potential process deviations. It is recommended that an RPA process analyst interview the process owner several times to uncover all the intricacies of the business process, and even observe the process owner execute the process in real-time to fully understand the depth and breadth of the process to be automated.
A new way of working
Organizations tend to be laser focused on the rollout and execution of RPA, with less thought given to the coexistence with bots.
As RPA adoption increases, an organization workforce becomes an increasingly hybrid mix of human and digital workers, which raises many questions that organizations may not have faced before:
- Do digital workers require user identities?
- What security controls and privileges are required for digital workers to access data and applications?
- What does this mean for audit and compliance?
- Who will oversee and manage the digital workforce?
- Who will monitor the new automated process to make sure it’s executed correctly?
- If the bot does not run correctly, who should the business reach out to?
- What higher value activities can we now address that we did not have time for in the past? How do I retrain human workers for these activities?
- What new and / or additional roles do we need to hire for?
The list goes on. Having a solid change management component as part of your RPA strategy instead of it being an afterthought will help your organization to achieve sustainable, long-term scale.
More than cost savings
When asking organizations about their objectives for deploying RPA solutions, ‘cost reduction’ is the first response that I usually hear. And it’s true – RPA can indeed provide enormous cost savings, especially in back office operations where repetitive, manual processes are more prevalent.
Alternatively, customers should look beyond the lens of cost takeout to achieve the full, transformational value of RPA. While cost reduction is a natural starting point, organizations often realize more strategic value drivers that RPA can bring to the business. Improved customer and employee experience, increased compliance and risk mitigation, and even business growth are just a few other outcomes realized by organizations today and where value from RPA can be quantified.
Conclusion
The potential RPA benefits for an organization are high and deploying low-complexity RPA bots is not particularly challenging. But as a long-term strategic imperative to significantly reduce costs, improve customer experience, or foster business growth, there are significant obstacles in capturing RPA’s full transformative value.
While it may be tempting to quickly deploy RPA to see quick wins, organizations should invest the time to set up their RPA teams for success and to ensure that RPA can succeed as a sustainable enterprise solution.
