Fact : your average Finance function will not be short of routine manual processes.  Allocating cash, reconciling bank accounts, matching intercompany balances, manipulating data – the list goes on.  Rules based processing is not the most value adding activity Finance does and, if you were to work out the cost of it across an organisation, it’s likely to be quite expensive. 

Robotic Process Automation (RPA) might only be a baby-step towards AI and Machine Learning, but it still has quite an allure.  Who wouldn’t want an application that can sit on top of your current systems and emulate rules based human activity at a fraction of the cost?  Pretty convincing, right? 

But not every Finance function is managing to take RPA past the proof of concept stage or to scale it effectively.  The top four reasons for this that I’m hearing are:

1.  Too many business rules have to be written into the robot from scratch.  The proof of concept works perfectly for Business Unit A, but Business Unit B is slightly different. To make the robot work, multiple rules have to be created and maintained for the entire organisation.  Anyone remember the pain of macros?

2.  Process standardisation becomes a pre-requisite to being able to scale RPA effectively.  Robots can be scaled quickly if they’re dealing with standard processes and data structures.  But this is often not the case as Finance is constrained by legacy systems.

3.  The business case doesn’t stack up as only a small part of someone’s roles can be automated.  Take an accountant for example, perhaps a small part of the month end is routine and can be automated, but there’s still the rest of the month.

4.  Building a robotics capability becomes too costly.  People are needed who know not just how to program robots, but also how to re-engineer processes and use the technology effectively to create efficiencies in the Finance function.  You need an accountant/ programmer/ process consultant all in one, which proves costly to develop or acquire.

From my experience, too often Finance is attempting to use RPA to plug the gaps in an old ERP system with outdated functionality.  Going back to the examples of routine finance processes above (allocating cash, matching etc.) all of this can be done by a modern ERP.  Robots don’t have to be programmed from scratch. 

Of course, the issue for most Finance functions is that getting the next ERP upgrade feels a long way off.  Software as a Service gets round this issue.  It also helps standardise processes, embed leading practices and Total Cost of Ownership is lower…

Lynne Sampson’s article below explains SaaS perfectly and is well worth a read for anyone who has more of a Finance than systems/ IT background and wants to understand the benefit it can bring.