Five keys to build success in RPA
“The first rule of any technology used in a business is that automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency” said Bill Gates. This is so relevant in the context of digital ambitions for most enterprises to leverage automation to the extent possible. Enterprises would like their skilled workers to get free from doing routine tasks to focus on innovation, creative problem solving, collaboration and deliver profitable growth. In a study by Lawless Research, highly automated companies are six times more likely to experience revenue growth of more than 15% versus companies with low automation. Close to 91% of companies believed that their skilled workers spent too much time on routine tasks. On the other hand, the volume of work is only believed to be increasing. As the world becomes more connected, this pace is only expected to grow.
Robotic Process Automation (RPA), typically used to automate routine and predictable processes is demonstrating significant potential to transform workplaces. In a study by APQC, the number of RPA projects double from 2017 to 2018 – with 14.9 average number of projects in 2018 versus 8.6 in 2017. Almost 50% of the cognitive projects were RPA. In most of these projects, 85-90% of the transactions can be completed with minimal to no human interaction.
The bots (codes on a server) impersonate the human activity and interact with multiple systems to conduct the process just like a human does. The typical journey in any RPA investment entails the following;
- Defining the portfolio of projects – It could be worthwhile to do a small consulting engagement to define a portfolio of projects including functions and processes that need automation. This is perhaps the most crucial step in the implementation cycle. To maximize returns from RPA, it is important to pick processes that take a lot of time, are routine in nature and can be clearly defined. Key factors in such decisions include aspects of scaling challenges when done manually, error rates, cycle time to execute the process as well as the time taken for resources to get trained and related attrition level challenges.
- Understanding the technology landscape and what type of technologies meet what kind of outcomes. For instance, a process automation for the procurement function may entail technologies like vendor portal, a standard full-blown invoice automation solution, OCR solutions, RPA solutions or a combination of these. Getting a deeper understanding of the solutions and their respective plays would be critical before beginning the vendor evaluation cycle.
- Selection of the right technology and partner: This step would entail definition of success criteria and then deciding the right technology vendor as there are numerous vendors right from boutique firms to generalists.
- Launching pilots before decision: A pilot would be a preferred approach over say a proof of concept in cases where the gap between the current and desired solution capabilities are not obvious. This is to ensure that the solution is indeed the right fit.
The evaluation process is not as simple and perfect as it sounds. Despite ensuring a robust evaluation, RPA projects could still not provide the desired results. On the other hand, when successful enterprises are shown to have reduced the cost per process by a factor of over 40%. RPA is no magic wand. Here are the top 5 keys to ensuring success in RPA:
- Selection of the activities to automate: This is perhaps one of the most crucial factors for success in RPA projects. If anything, this evaluation parameter needs to be assigned a much larger weight to influence the eventual outcome of any RPA investment. It is important to understand that RPA is not a substitute to automating a process-based technology system, but in fact a set of manual, rules-based activities. Some business activities are far more amenable to RPA. The RPA experts would sit with the process experts to map the activities targeted for automation. A general thumb rule would be to eliminate activities that have significant dependencies on factors or variables. It could be argued that simpler activities could be taken up first for automation since they deliver returns within a short timeframe. Very complex activities that require integration with multiple applications and far higher number of steps could come up later. Also, activities that require to be attended 24/7 or during post office hours could be better candidates. The activities should be definable, standard and constant over a period of time. Activities within the process that change more frequently may not be the right candidate. Typically, a lot of back office processes comes within the domain of RPA as they tend to be transactional and repetitive. RPA could be a good alternative to outsourcing, as the process could then be controlled internally in line with compliance requirements. Again, companies who are more mature in the automation journey will probably apply the technology in a far different way versus an organization who is just kickstarting the journey. A governance structure that focuses on the projects with maximum returns should be designed for on-going RPA projects.
- Optimizing the process before automation: Once the activities / process has been defined, it could be a good investment of time to quickly understand whether the process needs optimization before automation. There is no point in automating an already cumbersome process.
- Change Management and Stakeholder management: This is perhaps the most difficult part of any automation. Getting buy-in from key stakeholders and ensuring adoption is difficult in several enterprises where there is a lot of resistance to change. It is important to communicate the benefits with RPA to all parts of the organization. Employees who are wary that there could be significant job losses need to be assured that the robots are not taking away the jobs. Managing expectations of stakeholders to ensure that it is representative of realistic outcomes is equally crucial. The engagement must be strong both on the business and IT side for any RPA project and should have solid sponsorship from both business and IT.
- Establishment of a COE: Once some initial projects have seen success, establishment of an RPA COE is ideal to expand this value across the enterprise to cover multiple projects. The COE typically is just not about only building competence in technology but intended to drive business outcomes. As such, it may be productive to organize the COE around business objectives. In addition to driving the skills, tools, technologies and processes involving RPA, the COE also supports change management and faster adoption.
- Strong alignment between business and IT: While is a part of change management, this deserves a specific mention given its importance in the success of RPA initiatives. Since the IT organization is far better positioned to handle all the technical nuances especially with cloud-based applications, it is important that business and IT are strongly aligned to deliver outcomes. Once the RPA solution is functional, there are ongoing maintenance needs due to regulatory or other requirements requiring some changes in the process. This again requires close involvement between business and IT.
RPA is not so much a cost reduction play as much as being a play to quickly help enterprises grow in a non-linear way versus deploying trained headcount to meet such growth. RPA could be easy to configure and implement. Simple projects and numerous small tasks could be executed fast without entailing too much costs or timeline. RPA works with multiple back end applications without the need to re-architect those applications. When done well, it can bring in significant realizations of returns for enterprises.
RPA is also believed to lay the groundwork for more intelligent applications. As a side effect of doing ‘business as usual’ with RPA, the data that is collected is now going to be really valuable, because one can learn from it to derive insights, operate the process in a far more intelligent way and deliver smart applications. For enterprises, RPA could lay the foundation for AI even as the technology continues to evolve. With the increasing interest on automation with RPA, the day will probably not be far when the prediction of John Maynard Keynes comes true and humans will work an average of 15 hours per week and devote the rest of the time to creative and leisure activities.