5 Pre-requisites to strategic RPA
Any conversation today around automation would have one question in common:
“Should we invest in RPA for automation?”
In recent times, RPA’s adoption as an automation approach has become mainstream and many companies have started looking at RPA as one of their key strategic investment areas. As a sponsor of such an initiative, how can you increase the chance of a successful outcome? Based on experience from multiple automation programs, here is a five-point checklist that can act as a pre-requisite before embarking on implementation.
1. Have the goals been clearly defined?
Most teams have some idea of what to achieve through automation, but few have a clear quantitative goal defined. Like with any strategic initiative, without a well-defined goal, the stage is set for confusion with each stakeholder defining their own version of what must be achieved and this eventually impacts the outcomes of all stages of the implementation.
Before starting on an automation project/ program, it is essential for all stakeholders to agree upon the expected output and quantify them. The goal-setting can be done at different levels including organization level, division level or team level. Reduction in manual efforts, improvement of quality or turn-around time SLAs, scaling up the number of transactions etc. could be some of the parameters that could be quantified.
e.g. “Reduce manual effort in operations team by 40%”
2. Are the opportunities for automation identified?
Once the goals are defined, the next step is to identify what must be automated to meet the goals. This requires analyzing the organization’s processes and identifying areas that could be automated. For example, typically in an insurance operations team, a lot of time is spent reading through documents (quotes, policies, endorsements, claims etc.) and entering data into different systems (Broking, AMS, Claims Management etc.). At the operations team level, “Data Extraction from Documents” and “Data Entry in Systems” could be two automation opportunities that could help achieve the reduction in manual effort. With these 2 high-level opportunities, further break-down of process-wise opportunities can be done.
The output of the stage validates if the goals defined are realistic or need to be re-factored.
3. Is RPA the right tool for harnessing the opportunities?
RPA tools come with a host of features that suit many automation scenarios. Typical characteristics of an opportunity that is suitable for RPA are “Repetitive”, “UI Centric”, “Rules Based”, “Structured Data”, “Low Process Variance” and “Stable applications in the medium term”. If a process conforms to these rules, RPA tools out-of-the-box can quickly deliver automation. Understanding how many of the opportunities identified for automation are suitable for out-of-box RPA will determine the complexity of the RPA implementation. Though, the functionality of RPA tools can be extended through custom modules, identifying these custom modules upfront will enable the right estimation for the projects/ program.
Given that most organizations have other existing automation tools like workflow engines, evaluating RPA against building custom flows or applications on the existing stack has to be done. Added to this, an ROI assessment based on the size of the opportunity(s) and the cost involved in licensing, infrastructure, development and maintenance of the RPA stack would further build a business case for RPA.
4. Have other components to complement the RPA solution been evaluated?
Most processes have some amount of complexity that cannot be solved only through out-of-the-box RPA features. To expand the scope of automation to cover these areas and improve ROI requires development of custom components for integrations and business rules processing. Some of these could be components that plug into the RPA workflow while others could be pre/post-processing components to RPA. Two such areas which add most value are:
- Machine Learning — Complementing RPA with ML solutions enables expanding RPA to areas that have unstructured information or complex processing rules. Most often, this is the reason why such processes remain manual and are difficult to automate. For example, emails and documents are integral to many processes and are typically unstructured. Using ML approaches such as Natural Language Processing enables extracting information from such unstructured sources to feed the RPA workflow.
- Process Visualization and Analytics — Process automation goes hand in hand with process visibility. To ensure that automation is working as expected, business stakeholders need to be equipped with dashboards that indicate the health of the process, transaction volumes, activity of the bots and efficiency of the ML components. These dashboards ensure that any exceptions are addressed on time without leading to escalations. Further, analytics on this information can help identify next set of opportunities for automation.
5. Are the business users onboard with the automation strategy?
A great strategy on paper will still fail if it is not accepted by the business teams whose processes are being automated. An automation strategy document that showcases the goals, opportunities, automation approaches, ROI and road-map is a great way to rally support. It ensures that all key stakeholders are onboard with the strategy and provide the required support for the implementation. It also ensures that findings from this initial analysis act as guidelines for the automation program when the implementation team is brought in, team members change over time or the program expands to include additional business stakeholders and their teams.
Instead of starting the automation conversation with RPA as “the” solution, a strategic approach to evaluate the opportunities and identifying the right areas where RPA gives the best ROI and complementing this with other approaches such as Machine Learning will enable long-term benefits and a successful outcome.