Lessons learned from a successful Robotic Process Automation (RPA) Project

Robotic Process Automation (RPA) technology promises to deliver a disruption which can automate 40-90% of a business process. These are large claims but how do we know what is hype and what is reality?

We have recent experience in leading one of the most successful RPA implementations and there were a number of lessons learned which mitigated risk and helped to deliver benefits even faster than expected. Let’s explore these lessons using the five dimensions of business transformation.

  • If you are about to embark on an RPA journey, this article is for you!

  • If you are new to RPA, and would like to learn more about what it is, and the benefits it can deliver, click here.


Organisational design considerations

While building the business case, define the appropriate organisation design and target state operating model. Does the organisation want to build internal capability to utilise this tool or does the organisation want to buy or outsource the ongoing maintenance, support and operations?

Selecting the right mix of resources

When selecting an implementation partner or building an internal team, screen people for production implementation experience:

  • Which RPA tool do they have the most experience with?

  • Which other applications have they integrated? (e.g OCR tools)

  • Do they have experience adhering to coding standards?

  • Have they implemented a robust reporting mechanism to provide bot performance insights?

Avoid diluting team effectiveness by employing generalists. RPA requires process experts and skilled developers - make sure you get the right mix of both.

Train people early

Up-skill and ensure your process owners and subject mater experts are trained early in RPA capability (preferably before kick-off) so that maximum potential is reached during the critical process design phase. Don’t forget the power of RPA mixed with other tools like Intelligent OCR.

Keep in mind that robots can work 24x7 - nights, weekends, public holidays, etc. This is a significant benefit particularly when targeting processes that impact cash flow (e.g. billing).


Tool selection

Do your research and define your selection criteria. Go with the tool that is fit-for-purpose for your organisation and desired operating model (leveraging existing skill-sets in the organisation will be beneficial). Most of the tools have a trade-off between functional capability, user experience, ease of use and back-end power. Make sure you validate the claims of software vendors, and don’t fall victim to overselling.

Coding standards

RPA implementation programs generally have significant time constraints, requiring multiple processes to be developed in parallel. This means there will be multiple developers working in parallel all with their individual styles of development. Establish clear coding standards and guidelines before development starts. IEEE offers standards for Process Automation, and the software vendor will usually offer coding guidelines that work best for their tool.


Process Owner & Stakeholder Sign-Off

Make sure that your Process Owners are actively participating in the process design phase and are aware that signing off on the target state process design has real implications for their ongoing operations. All steps that are process mapped will be developed and any missing steps can be expensive to redesign and redevelop.

Ensure that IT are involved early in the discussion as it will be critical to align with Enterprise Architecture. IT should support the business in the delivery of appropriate infrastructure that is designed for performance and is maintainable post project.


Business Case and Benefits Realisation.

Effective projects are clear on their objectives from the business case and maintain a laser focus on benefits realisation. RPA is no different. Clearly define the benefits in the business case and remember RPA should be delivering high ROI multipliers of 4x or above, 7 or 8 figure NPV values, and payback periods measured in 12 to 24 months, not 3 to 5+ years like traditional technology projects.

Although the benefits can be realised faster than you might expect, don’t expect perfection on Day 1, there will be some issues to work through, but our experience has shown these to be relatively minor compared to the value of the overall improvement.

Operational control

Defining how your human workforce interacts with your digital workforce (and vice-versa) is key for operational control and effective change management. Define accountabilities and responsibilities and align these with existing operational risk frameworks. Map the change control procedure for business and IT governance frameworks, monitor the bot performance and audit bot permissions.


Organisational Pace

Understand RPA’s ability to deliver fast value by skipping expensive and often wasteful proof-of-concept stages, and go straight to a Production Pilot. This will build trust with your stakeholders and reinforce that RPA performs exactly as well as you develop and train the system to perform.


Ensure there is strong executive sponsorship for the initiative and support for team members. Sponsor messages should be loud and often to reinforce the need for change and build additional awareness across the enterprise. RPA is often the first dip in the AI pool and will require strong leadership to take the next steps in to Intelligent Automation and Artificial Intelligence.


Ensure that the team values are set early such as low politics, collaboration, merit based decision making and high performance.

Change management

Change management is a vital component of a successful RPA program. These kinds of disruptive technology projects are not your usual system, application or infrastructure upgrade. These projects have a direct and profound effect on people. The very nature of the way your people work will be changed to focus more on exceptions and less of the repetitive tasks. Make sure the team is trained on the new way of doing things well in advance of the first day of operations in production.

That’s it, for now. Over the next few weeks we will be exploring Intelligent Automation as part of our Applied AI insight series.

Have you implemented a large scale RPA or Intelligent Automation Program recently? What lessons can you share from your experience? Comment below.