Tricentis sponsored this post.
Companies are increasingly implementing robotic process automation (RPA) as part of their automation strategy, yet scaling RPA remains a major hurdle. Numerous research studies have shown that scaling an RPA initiative is more difficult and costly than expected. By some estimates, only 13% of RPA projects advance beyond the pilot stage, and the majority of organizations implementing RPA have fewer than 10 bots in production.
Something is clearly wrong — but how do we fix it?
To better understand the barriers to scale and help organizations avoid dangerous missteps in their RPA journey, Tricentis commissioned Forrester to research what RPA programs really look like on the ground and identify what challenges firms face related to scaling those programs. The results are captured in the Forrester report Barriers and Best Practices for Scaling RPA.
Key RPA Scalability Research Findings
Key findings from the Forrester research project include:
- Firms can’t seem to master RPA resiliency — and that impacts costs. Less than one in five firms are very effective at resilient automation, and firms that struggle with resiliency are also four times more likely to say they are very ineffective at controlling costs associated with RPA.
- Advanced programming skills are usually required. Although RPA often promises to make business users “citizen developers,” 79% of firms report that their RPA programs require advanced programming skills. On a scale of one (no technical skills are required) to five (advanced programming skills are needed), the average score is 4.1.
- Bots often break, placing a heavy load on resources. Forty-five percent of firms deal with bot breakage on a weekly basis or more often, and in the meantime, customer experience is impacted. Employees also have to take on additional manual tasks that would otherwise be automated.
- Scripting permeates today’s RPA implementations. Virtually all respondents (99%) say their organizations’ bot logic requires some scripting, with a weighted average of 42% of bot logic expressed within a script.
- Firms using scripting-based RPA solutions experience more pain from broken bots. High-script firms were 1.3 times more likely to experience impacted customer service from broken bots, and 1.5 times more likely to have impacts to cost or revenue, e.g., from delayed transactions.
RPA Leaders vs. Laggards: What Leaders Do Differently
The research also analyzed the data from RPA “leaders” versus RPA “laggards.” Some of the key differentiators they discovered were those leaders:
- Minimize scripting (11.4% weighted average compared to 45.9% weighted average).
- Prioritize RPA tools that reduce maintenance and deployment times (based on their top 3 tool capabilities).
- Leverage their mastery of resilient test automation for RPA (firms with effective test automation are better at controlling costs and keeping bot maintenance issues in check)
Recommendations for Scaling RPA
The report concludes with the following recommendations for getting ahead of the challenges identified in this research:
Develop a proactive approach for bot resiliency. Application changes and system issues are leading causes of bot breakage, affecting revenue and customer experience and adding to staff maintenance costs. Work cross-functionally to communicate pending app upgrades and work with infrastructure and operations professionals to implement the latest failure analysis and self-healing technology.
Include customer experience (CX) and revenue in your hunt for operational efficiency. Cost reduction through the integration of legacy apps on a processor’s desktop is the typical RPA use case. This is all good, but RPA can also generate new revenue and improve CX. Both debt collection and reduced cycle time for an order-to-cash process are fertile areas for revenue. CX can have more dramatic results.
Build an assessment model to guide RPA process selection. RPA plugs gaps in legacy systems, but it will sometimes end up delaying needed system modernization. Some processes will benefit more from new digital approaches, not from a patchwork of RPA robots doing the same old process with slightly less labor. Apps with a less than stable history are not good candidates for automation.
Start centrally but plan to federate responsibility to the business. Formalize the operating model early in the RPA journey. It’s acceptable to tune operations in a small centralized team, but plan to spread out outlined functions to individual business units. Develop a joint understanding between business and technology teams before automation projects get out of control. This will drive a quicker set of results and avoid a host of potential RPA issues.
Build the operating model around automation — not a specific RPA tool. Forrester has identified 13 AI technology building blocks that can add intelligence to the digital workforce of the future. In addition, you may end up with a couple of RPA tools that specialize in a process domain such as finance, contact center, test and automation, or IT service management. Building operating models tightly around a single RPA vendor is a poor long-term play. Instead, follow a broader strategy that focuses on automation beyond — but including — RPA.
Keep change management initiatives front and center. Internal users and immediate managers, despite stated desires, have a great fear of change as well as heightened anxiety about robotic solutions replacing them. Leading companies will allocate required resources that make change the norm, not a feared outcome.
Feature image by from Pixabay.