Having already covered how RPA can improve business value and RPA adoption best practices, we are now ready to get down to work and look into RPA implementation.
A typical RPA transformation journey follows a normal thread:
As we can see, reaching the digital transformation phase takes time and previous experience with basic and/-or advanced RPA. That is why 76% of companies are only halfway into the RPA implementation journey. They are part of the so-called Bot 0.0 and Bot 1.0 companies, having experienced RPA in production for up to 1 year.
Best practices for a successful RPA implementation journey
First things first: a company looking to adopt RPA should undergo a self-assessment, to see if it’s prepared for implementation. Harsh and restrictive politics within a company are never a good sign, because some vital communication channels could be closed. Lack of communication and poor change management are the reason why, statistically speaking, almost 50% of RPA projects fail, according to EY.
Whether you choose to venture in this transformational journey on your own, or you choose an RPA expert to partner up with, here’s a solid set of implementation rules and best practices for you to follow when adopting RPA:
1. Start out with solid research and establish a project plan
Your analysis should include everything from choosing the right RPA technology, to creating solid business cases, building up a relevant proof of concept, developing viable ROI metrics etc.
It's recommended to start off in this journey with an expert RPA consultant by your side, to help you define the project’s scope, strategies, expectations, available resources, key deliverables, timelines, possible risks and solutions to counteract them, and so on. While few will know your company as well as you do, a fresh, experienced set of eyes might be able to see where you should be heading more clearly.
2. Target business areas best suited for RPA
How to recognize these? Aim for business processes with little to no human interaction in the automated version, where a faster execution time is possible, improved data quality is guaranteed, and increased employee satisfaction is a probable, if not certain outcome. Develop a robust change management strategy, as the majority of processes have dynamic variations which need to be thoroughly documented and implemented when these processes undergo changes.
3. Benchmark process variations
Before the commencement of the project, ensure you have the correct metrics about process variations, frequency, error rates etc. – any values that characterize the process you’re looking to automate.
4. Abandon the process if it’s too complex
A complex process is by no means something you should start your RPA venture with. But if, for whatever reason, somewhere down the road, you find yourself investing too much time, money and energy in a process that is too complex, it’s best to leave it as it is. It might never be suitable for automation or it will be a better candidate sometime in the future.
5. Develop sustainable, simple and modular software robots
Treat the robots as reusable objects. Building them upon a simple structure will help minimize the number of failure points and facilitate updates and tests. Robots automatically stop if they encounter process changes they don’t know how to handle. And since you’re not looking to automate in order to have things move slower than they normally would, it’s best to keep the robot infrastructure simple and ease to refresh.
6. Develop in-house skills alongside robots
In order to keep a healthy pace for your RPA journey, you need to invest in training in-house development resources. This is another one of the many benefits of working with an experienced RPA partner – they will create a winning dynamic for you, by working alongside your internal team and training them accordingly.
7. Ensure data security and engage IT
In order to ensure a stable and secure automated process execution, there needs to be close collaboration with the IT department, as they will be held responsible for the effects generated by implementing this modern IT solution.
8. Periodically test implementations
Set milestones and constantly test the developed solution to uncover weaknesses and involve business users in UATs. It is very important to develop a solid testing phase and create robots capable of achieving the agreed SLAs.
9. Develop a feasible rollout plan
Work with consecutive roll-outs: first in the test environment, then gradual deployments in the production environment.
10. Measure the outcomes
Develop a framework to measure performance across deployed robots, process variations etc. – this way you’ll know where there is room for improvement.
11. Keep yourself open for future advances or challenges
RPA’s complexities will soon reach new levels, such as self-analyzing process improvements or self-building bots.
Be aware and beware of potential RPA implementation pitfalls
The most common RPA implementation pitfalls range from a lack in management support, to poor planning, to the selection of non-RPA use cases, and everything else in between presented in the image below:
Most of the road bumps you’ll come across on your way to RPA excellence, are related to exception handling, business exceptions or system exceptions. The robot needs to correctly apply a failure configuration (safe exit the process) or restart configuration to overcome these bumps.
Error handling goes hand in hand with maintaining a functional automation – when a system is temporarily down the robot should re-trigger the failed cases.
All of these pitfalls can be avoided through effective communication between stakeholders, creating and following a solid project plan and scope, increasing awareness regarding RPA and its capabilities, and defining the accountable persons and reliable metrics for the project’s outcomes.
An article written by our colleague, Marius Ungur
Stay tuned for our next article to learn more about RPA vendors, industry insights and cognitive automation.