As automation technologies like robotic process automation (RPA) are advancing, they are also becoming easier to deploy. But the challenges remain, and the complexity of the technology shouldn’t be underestimated, especially as we progress towards more elaborate, AI-enabled forms of automation. RPA is a technology that is easy to use, but getting sustainable results takes time, and so does an RPA implementation.
It may come as second-nature eventually, but it is likely to be accompanied by learning curves, obstacles, and stumbling blocks for companies and their employees. During our #UiPathForward global event series in NY, London, Bengaluru and Tokyo, we welcomed more than 4000 customers, partners, and RPA developers all eager to learn from each other’s experience with RPA. Sophisticated organizations like GE, HP, Allianz, Generali, SMBC, Dentsu, VMWare, Deloitte, and KPMG were there to share their stories. They are all at different stages of maturity in their deployments, therefore the wealth and diversity of stories were enlightening. It was worth distilling them to reveal the main lessons to be learned. These should help new adopters sail more easily through the sometimes muddy waters of RPA implementation.
Start small, but think big
"Don't go in trying to capture the big, complex automations. Start with the small ones that may not get you the business value right away, but start with those in order to train your employees how to do this and start exposing the rest of the business to what RPA is and how this could actually help them." Mike Macdonald, Associate Director Research & Technology, Symcor
The saying Don’t bite off more than you can chew holds true with RPA deployment, since initially automating complex process will only create delays. Repetitive tasks that don’t require judgment (e.g. copy-paste jobs, data migration) should be automated first. And the first phase of implementation should involve running a Proof of Concept or pilot to demonstrate both the usability and potential of RPA within the company. Only when productivity and efficiency benchmarks are attained should bigger projects be tackled. At the same time, small-scale, conservative planning is not the best approach for achieving deep technological transformation with large-scale RPA deployments. To eventually achieve large-scale automation and to extract the greatest value from RPA, it’s increasingly important to plan holistically and end-to-end right from the start in order to eventually achieve large-scale automation.
Use an automation roadmap, and don’t rush the clock
RPA deployments can easily slump or even fail without preemptive planning, and adopting RPA in an unorganized, complacent fashion can lead to a number of challenges. As such, it is crucial to establish an automation roadmap that details the RPA software purchase as well as its implementation process and long-term prospects within a company. This strategy should also be clear about how RPA will unfold within an enterprise: When and to what processes should RPA be applied? What are the benefits, risks, and goals of the RPA project? Will RPA be used as a permanent solution? Who will be responsible for supervising the automation? The automation roadmap must also outline the time that will be necessary for each stage of the implementation process.
"In most instances, we haven't completed the process improvement before we implemented RPA. We were doing the two together, and what we found is that that's a more efficient and a more modern strategy than just implementing Lean Six Sigma on its own." Darryl Neff, Group Lead Lean Six Sigma & Robotic Automation, Generali
Keep in mind that RPA deployment shouldn’t be rushed, and ample time should be allocated to solution testing.
Establish an RPA Center of Excellence to “build for scale”
In addition to relying on a well-developed strategy, establishing a Center of Excellence (CoE) is the best practice to create a foundation for scaling the RPA initiative. It’s especially important to set up a CoE and establish the accompanying operational standards early in the implementation process, in order to avoid reconfigurations of unstable, unsustainable solutions with future automation. Employees within the CoE should oversee and drive the implementation, roll-out, and long-term operation of the RPA initiative. In this regard, the CoE will help prioritize processes with the potential for automation, ensure that quality and compliance standards are met, and allow for coordinated communication with the relevant stakeholders.
Curate an RPA team and develop competencies among staff members
Though many RPA solutions are designed to be highly intuitive with graphical, easy-to-use interfaces, a lack of adequately trained automation specialists or appropriately skilled process subject matter experts can risk the success of an RPA implementation project. And due to a supply-demand gap for RPA specialists, it is often difficult for enterprises to acquire the needed resources externally.
"What I am seeing more of is clients wanting to take more of a bigger leap, i.e. not do the really small proof of concepts, but move straight to doing pilots and starting them quite fast." Helena Clennell, Partner within PwC's Finance Effectiveness Practice
The maximum benefits of automation cannot, however, be attained without the internal resources and talent to support RPA’s technological transformation. In order to reduce implementation risk, it is crucial to establish these skills in-house. Companies looking to deploy RPA should carefully evaluate a vendor’s product training tools (e.g. webinars, videos, resource libraries) as well create focused trainings through its CoE for each role on the RPA team: business analysts, developers, controllers, service support figures, and project managers.
Ensure streamlined internal communication and change management
Communication is key. Between members of the implementation team, but also with the rest of the company, throughout the entirety of the RPA implementation process. Change management is crucial from the start, in order to communicate to employees their changed roles and to avoid resistance to the initiative. This also allows employees to embrace the benefits of RPA, rather than fearing the technology will replace their roles.
However, communication silos can also emerge later on as individuals within an organization feel cut-off, making it difficult to exchange insights and make gains. Creating a routine communication schedule will reduce the amount of uncertainty and ensure clarity, and employees will be aware of the implementation process and what the implementation of RPA entails for them individually.
Don’t forget the business-IT partnership
As a result of lacking collaboration between business and IT departments, enterprises often face difficulties in implementing RPA within the planned time-frame. But maintaining consistent communication and strategic alignment between sides — already in the earliest stages of choosing an RPA software — is especially important in making sure that implementation goes as planned and that issues are dealt with seamlessly.
“We learned a lot about setting up the IT infrastructure for the RPA implementation. We now have it in place, and we’re working with our group IT colleagues to make it even more robust for the future.” Mark Jenkins, Regional Head of Organisation, Euler Hermes
Long-term, the IT team must be responsible for developing, testing, and maintaining automation, all while working alongside the business operations team in managing strategic goals and expectations. Managing such organizational change is never easy and always takes time, but business-IT teamwork will ensure that a growth plan is elaborated to sustain RPA expertise beyond the earliest stages of implementation.
The main takeaways
It’s easy to see that RPA is not the end-all, be-all, cure-all to operational woes. That being said, RPA also delivers huge benefits in efficiency, productivity, cost reduction, and risk avoidance when the necessary efforts are investing and planning is executed. Consider these “lessons learned” a roadmap for what can go wrong and what to do right through the major stages of RPA implementation: from considering RPA to a proof of concept or pilot test, to the final stages of implementation and RPA scalability.