Mistakes to avoid when implementing RPA


Implementing Robotic Process Automation isn't a simple process, but it's important to make it as seamless as possible to ensure maximum benefit early on. Here are 8 common mistakes that businesses should avoid when adopting RPA!


  1. Not choosing the right processes

It might be tempting to automate small, simple tasks that nobody wants to complete. While this may be useful, the most beneficial tasks to automate are those that use significant human resources. End-to-end automation of these processes are where you will see the best ROI from using RPA.

Example of these processes include:

  • Credit decisions in banking
  • Underwriting decisions in insurance
  • Processing account changes
  • Getting customers set up across a number of industries

To ensure the benefits from your RPA project, make sure it can be measured with one of the following KPIs:

  • Generating revenue
  • Reducing costs
  • Improving quality
  • Positive impact of audit, risk and compliance

To define which processes will be the most efficient to automate, it's important to have a defined vision. Businesses can safely select processes which involve any of the following criteria:

  • Rules-driven: processes that are consistent
  • Data intensive: tasks that involve using large data sets
  • Repetitive and manual
  • Electronic trigger: processes that start when having received electronic data files
  • Tasks involving a manual calculation of results
  • Out of hours jobs: tasks that require employees to work during anti-social hours or that would benefit from being carried out 24/7
  • High error rates: these tasks are often paper-based or able to be made in error by human interaction
  • High compliance: tasks which need to be audited for regulatory compliances
  • Validations: tasks which must be validated by multiple systems
  • Multi-step processes


  1. Automating broken processes

In most cases, businesses will use RPA to automate existing processes. However, make sure that processes are streamlined before automating them.  This will ensure more efficient decision logic for RPA bots, shorter RPA testing cycles and easier maintenance of automation solutions.

E.g. if you want to automate the onboarding process for new customers but there are several different ways your business can onboard them, figure out which is the most efficient and automate it. You'll probably find you can discard some, if not all, of the alternative methods.


  1. Automating all processes at once

There are varying opinions on this point, but generally speaking, it's better to implement RPA slowly rather than trying to change too much too soon. If RPA is new to your business, it's effective to start small to build confidence and support for the change. It's recommended to phase in RPA development by implementing a trial which you can use to learn and scale.


  1. Not investing enough time to test your bots

Taking the time to programme your bots properly so that they can handle each task properly will save you a lot of time later on. If bots are programmed to carry out tasks correctly, staff should only have to deal with exceptional cases that the bots cannot handle. To programme bots best, you need to carry our extensive testing to make sure that bots understand what they're meant to be doing.

No doubt when testing you'll come across snags and scenarios that can't be anticipated. Make sure you allow sufficient time during testing to accommodate these. 


  1. Not showing benefits early on

To ensure that leadership fully back the adoption of RPA, make the benefits visible to them as early on in the process as possible. Within about 3 months you should be able to show benefits to leadership. Demonstrating these benefits will prove its effectiveness but also help get support to implement RPA in other processes.


  1. Failing to scale up

It's one thing to set up one RPA bot for one process. But RPA is the most efficient when deployed across multiple processes. The less monotonous, manual tasks that employees need to carry out, the more time they'll have to use their expertise in more important jobs. Scaling up means to implement an RPA Centre of Excellence. These handle tasks such as

  • Training new RPA users
  • Overseeing RPA tech infrastructure
  • Owning relationships with RPA providers
  • Deciding which processes to be automated.


  1. Having a narrow focus on cost-reduction, without a holistic perspective

RPA is an end-to-end change program, not just an isolated automation process. The bigger picture should include other technological developments like AI, natural language processing and analytics with an end product of digital transformation. Make sure you start your RPA journey with a long-term business case which is specific in terms of goals, but flexible enough to allow the adoption of future innovation.