Following my talk at the RPA Summit in Sydney I received many requests to share more widely my insights from building and scaling RPA capabilities in large organisations. Here are the first two insights. Keep watching this space as I will be publishing more insights on this topic over the next couple of weeks.
1. A pilot project, a PoC and a prototype don't make an RPA capability
You may have successfully implemented a pilot RPA project or you may have a desktop bot running that is saving you every week several hours of mundane work. Congratulations, that is the important first step on your RPA journey, but let's be honest, a few standalone RPA solutions are not equal to a scalable capability that can be readily deployed in every business unit of your organisation. At this stage you may think "What's all the fuss about, this was way easier than I thought." True, and that's the beauty of RPA, you can very quickly achieve a reasonable result. You might even be tempted to go to your executives and promise them massive cost savings in every area of your business. But I'd urge you to be more cautious and not to overpromise on the net savings and especially on timeframes.
RPA pilot projects and proof of concepts usually don't care a great deal about compliance and legal concerns, information security, interdependencies and integration, prioritisation and resource constraints, operational issues and SLAs, change management and impacts on human resources, production support and maintenance cost, Disaster recovery and BCP, scalability and load balancing, exception handling, reusability and documentation, and, and, and .... prototypes are here to quickly assess the potential of the new technology and prove its feasibility. But cascading RPA solutions across your entire organisation will require to develop standards and frameworks and put an appropriate governance structure in place BEFORE you can deploy your bots in a repeatable and effective fashion.
2. Choose wisely which business processes to automate
After you have acquired your RPA tool and successfully implemented a pilot project, the word of the new technology will spread like wildfire and business leaders from all corners of your organisation will flock into your office hoping to be next in line to benefit from the "rapid cost savings" promised by RPA vendors and consultancy firms alike.
Overwhelmed by the seemingly endless number of business processes in your organisation you need a framework to assess quickly and reliably which business processes are most suitable for Robotic Process Automation. Fundamentally, when assessing an automation opportunity you should ask yourself three essential questions:
Can we do this (feasibility)
Should we do this (viability)
Do we want this (desirability)
Only if all three questions can be answered with a confident 'YES' you should pursue this opportunity. In order to thoroughly answer these questions you may need to consult with a number of technology, business and administrative functions within your organisation. Developing a questionnaire or assessment guidelines will not only help to standardize the assessment process but also make it more efficient. In my experience, the extra time spend at this stage of the RPA project usually outweighs the time and resources wasted when discovering half-way through the project that automating the business process is either not feasible or not viable or not desirable.