Robotic Process Automation (RPA) is the automation of business processes across the enterprise using one or more software robots. It mimics human behaviour (copies the actions of human user performing high volume, repetitive and rules-based tasks) with no change to the existing, underlying technology infrastructure.
With RPA there is no coding, just configuration of workflow to orchestrate existing assets. Therefore, the underlying systems will not change, and the automated process will only perform the same functions as the human workforce.
While RPA is a much wider term which could involve automation of physical (e.g. mechanical), as well as software, processes using robots, the focus here is on Robotic “Software” Process Automation. Any references to robot(s) is limited to software robot(s).
RPA has two main distinctive characteristics. The first is that RPA software requires no back-end integration with existing systems or database, because it uses the existing logic of applications and interacts with them as a human user would.
It’s also relatively easy to configure and, instead of programming a piece of software, it resembles the modelling of processes. Therefore, tech-savvy employees can be trained in only a few weeks to configure RPA, whereas specialised, and thus more expensive, knowledge is required for traditional process automation solutions.
Why RPA is important
In addition to ease of integration and implementation, RPA can potentially make business processes more efficient. Repeatable process steps can be executed faster, more accurately and cost-effectively than by traditional human workers.
By leveraging RPA, workers can focus on more strategic tasks while customer service teams can be more efficient. RPA can also improve the customer experience and potentially generate new revenue.
There are also compliance benefits, as the audit trail is always maintained, accuracy of data is improved, and human contact with sensitive data can be reduced. This, in turn, can lead to an improved customer experience, operational excellence, and front-end/sales enablement execution.
Decide what to automate
Before you decide to automate, it is important to understand in advance the complexity of the processes.
Rules-based transactions are prime candidates for automation, but if a process requires a lot of decision-making, implementing RPA may not be effective. Humans can make decisions by learning patterns or based on common sense, while RPA cannot yet mimic this human ability.
There are workarounds for these, but this will result in an increase in process complexity that can cause RPA software to require more effort and become more costly to configure, than continuing to use a human to perform the same task. For example, a loan application process is a rules-based transaction, but still requires decision-making using common sense, so implementing RPA might not be effective.
However, if the process is rules-based and it looks for patterns and makes predictions, then it can be a candidate for automation. An example is an advisor’s process to support the financial planning needs of customers by applying statistical algorithms to determine the best returns on their investment portfolio.
Plan your approach
You need to think up-front about what your vision is, since you don’t want to by-pass this. “Vision” may seem like a wishy-washy term, but it’s key to have that in place, as once you’ve gone through the 1st 1-2 pilot processes, you get burnt without it.
This is also the time to think about organisational structure, governance (bringing both the business and IT together), and capabilities. Schedule sessions to discuss functional and technical designs to make sure business and technology are on the same page.
Before making the decision to implement RPA, it’s important to hold “design sessions” that take into account the following key functional and technical areas:
- From a functional perspective, this includes mapping out business processes, gathering requirements and information from end users, collecting information on the frequency of process execution, and documenting process outputs.
- From a technical perspective, this involves securing application availability in advance for robot development and testing.
- Establishing access to the environments, especially if the underlying application is a third-party application. Identifying any differences between test and production environments that need to be accounted for in go-live, is also vital.
- Security is one of the most important factors to consider. How safe are the robots compared to humans? After all, robots/RPA tools are also a piece of software, so the tighter the security controls the better. You also need to understand how security patching is applied to environments, as this can impact your setup post go-live.
Start small and Agile
Being Agile can also mean a willingness to Fail Fast, or trying something, getting fast feedback and adapting. By innovating at a fast pace, Fail Fast allows a business to avoid big failures and enhances its rate of success.
It’s important to understand that failure must be smart, and that you need to learn from it fast. This is also key to getting wins on the board quickly.
The other point I’d make here is to manage expectations up front. RPA is very environmentally dependent – it can get thrown out by changes in Chrome browsers or Security patches – though these are easily resolved.
This implies it will need tuning in Production. This is not like your standard software development project where there may be no tuning or configuration in production.
Role of RPA in testing
RPA implementations need to be tested. They are designed to provide a payback to the business, and you need to verify that it will deliver.
With the increasing demand for better quality, the importance of testing cannot be stressed enough. It's crucial to check that business processes run across all scenarios and configurations.
Testing can't be limited to verifying business process. You need to:
- verify data inputs and outputs,
- check for scaling and concurrency issues,
- ensure the robot handles exceptions properly.
You may need to re-engineer the end-to-end processes, which includes the upstream and downstream process. It also means that change management will also be required and needs to be identified before go-live.
Modifications to existing processes may vary. Factors such as data inputs required, pre-existing data standardisation, number of underlying systems that the automation will touch, and flexibility in the sequencing of tasks can influence the amount of upfront work required before configuration and development begins.
Issues regarding deployment of software or IT changes in regards to technical and people change. Change management blockers, such as gaining access to system environments, getting security approvals and coordinating with the relevant stakeholders, are usually faced in the initial assessment stage.
For seamless RPA implementation, it is important to manage the expectations of the project team by establishing a communication schedule that outlines the reasons for the process re-design and the actions required by the team.
Change in mindset
The implementation of a large-scale RPA program is a complex and challenging process. RPA is a disruptive technology that changes the way people work.
In addition to the fear of change in general, there is a fear that robots will replace a human workforce. It’s important to emphasise that, while software robots start performing routine repetitive tasks, workers can focus on more creative and higher-value tasks.
Executive team support is critical to success of RPA projects, as an RPA implementation can be stressful for employees. It’s important that employees are informed about what are the expected business benefits, impacts and limitations of RPA, and what the changes in roles and responsibilities are.
Educating and training staff in the new skillsets and tools can support the business and process changes and help with the adoption of the new technology. You need to ensure that the communication and education is targeted to the users and stakeholders for each process. I.e. target the team affected, as well as wider comms.
Change management is critical throughout the lifecycle of an RPA project to ensure that employees are properly prepared to take on new tasks that come with process re-design and the upcoming robot deployment. Given the magnitude of change required by your teams, RPA requires a robust change management plan before go-live.
The other common pitfalls of RPA can be found in standard software development or COTS projects:
- Lack of time commitment from local team – very easy to get bogged down in day-to-day activities or BAU.
- Lack of leadership buy-in.
- Lack of IT ownership. Whilst RPA should be business-led and IT-enabled, you still need clear ownership of IT’s activities.
- Unclear responsibilities.
- Choosing a process with insignificant business impact.
- Choosing a too complex process.
- Choosing a process where better custom solutions exist.
- Lack of focus in process selection.
- Striving for end-to-end automation when it is not cost-effective.
- Choosing a solution that requires intensive programming, which leads to technical pitfalls.
- Maintenance resulting from post-implementation pitfalls.
RPA is now used for enterprise-wide deployments, and the reasons above should give an indication why this is the case. It is predicted that the increasing demand for automation of business processes will continue to drive RPA market growth.
It is important to remember that for RPA to succeed, it must be business-led with support from IT. Change management is also critical, ensuring that people embrace the change rather than shy away from it.