Skip to content

Robotic Process Automation (RPA) seems to be a silver bullet. But as any bullet, not used appropriately, it can cause major damages in the end.

In the era of "digital transformation", Time to Market is critical and business cannot wait for heavy projects. "Vintage" IT systems are not able to evolve anymore, or fast enough, and now digital transformation pressure which was upon CIOs is flooding overall business.  
Robotic Process Automation (RPA) is becoming the new silver bullet to all the issues business has with IT. But the winner might not be the one we think.

RPA, the shovel seller of digital transformation  

RPA is a hype topic

CxOs are more and more interested in Robotic Process Automation. And of course consulting and IT companies are more and more interested to sell projects around these.

(Nearly) All big IS&T or consulting companies are pushing this topic forward. As examples : Capgemini claiming to have "major successes", Accenture claiming to be "the leader", Cognizant also claiming to be "leader in process automation", Deloitte have their own "approach", And EY views it as "a massive catalyst for adding value".... 

Research companies plan a continuous growth of the automation and associated services market size. These can become the new El Dorado :

Some voices, like this post from Syver Anmarkrud try to be a bit more pragmatic, but for how long?

It can indeed bring real benefits

Automation can bring major benefits when talking about compliance or tasks execution consistency. However, most of people consider that IS&T systems are a way to generate more  unemployment. A lot of projects have been run in the past to reduce labor costs and consulting just created the expression "reoriented on higher value added tasks" to avoid saying "risks for your jobs". 

The fear is real and shall not be denigrated. The technology is not the issue in the end but how we use it. One of my past jobs experience led me to start implementing AI within insurance front to back processes. All the employees involved, from the agents to the legal experts, were struggling in repetitive tasks and complaining not having time to fulfill their commitment to their customers. Top management even added more employees, increasing even more administrative and repetitive tasks. In the end, the project was aborted because HR representatives feared jobs losses. 

Clerks working in 1937
Removing repetitive tasks from employees

Digital transformation shall definitely be driven jointly between business, IT and employees (and their representatives). Automation can bring high value even, for real, for employees as long as cost cutting is not the only driver. Appropriately used, it can increase not only productivity but also employees morale.

But...

... Implementing automation is not always a success

Human Business Process Management jobs are far from disappearing. Robotic process automation failure rate is 30-50%, said EXL CEO Rohit Kapoor early this year. Processes that are being automated can be quite complex. Even one of the major RPA software editor (UIPath) clearly acknowledges that a significant part of these projects fail, based on an EY report. However, some reasons for failure in this report can be challenged :

  • "Not considering RPA as business-led, as opposed to IT led" : True. Whatever the project, it should never be IT for IT. However, Agile approaches emerged partly because IT and business progressively built a Chinese wall between them. There is no such thing anymore as "business led" or "IT led". IT is the heart of transformation, whatever happens. Projects are based on business drivers and involve both business and IT at the same level. 
  • "Applying traditional delivery methodologies" : There is no such thing anymore as a "traditional delivery methodology". But whatever the approach, project delivery foundations shall always be applied : Don't rush before having planned and designed a little. Don't remove deliverables (they are the only key to maintainability). And don't forget testing!
  • "Assuming RPA is all that’s needed to achieve a great ROI" : True. If inbound documents or processes are massively paper based, as mentioned in EY post, achieving not a "great" but even a significant RoI will be impossible. But keep in mind that RPA is a way to quickly bypass IT issues, it does not fix them!
  • "Underestimating what happens after processes have been automated" : True. We can never say, specifically with this kind of solution, that a process is automated and that's over. I'm not so sure that only putting in place a dedicated maintenance team can be the only success factor...

Current RPA instantiations take the assumption that time is frozen!

The only constant in the technology industry is change

Marc Benioff

Robotic Process Automation is based on two observations : 

  1. All front to back, back-office and support processes within a company are dealing with routines, not directly linked to intrinsic process complexity but to tools put in place (IT or not)
  2. In the case of IT issues, processes could be simplified but evolution to core business tools costs a lot, is risky and requires a lot of change management on human resources. 

Lets add to this, when going a step further with "Intelligent Process Automation" (IPA) or "Cognitive Automation", the assumption that AI can bring value anywhere, as a consequence, also in the automation area. In that case, AI is seen only as a synonym to machine learning and to some extent to Natural Language Generation. 

>Read more on Artificial intelligence types and use

As a consequence, RPA is at first a quick win to bypass the current state of IS system. Taking into account all the ongoing evolution regarding business, offers, processes and front-end IS systems, even if efficient, it makes me think of this :

Implementing RPA while really doing digital transformation

Being able to anticipate the risk of systems evolution is already great. Knowing that putting in place Robotic Automation can give you some time, but is not the final answer, is even better. 

Do not reduce your costs but save to reinvest!

Robotic Automation can actually have great power to bring short term cost reductions or productivity increase. But putting in place this kind of tool can quickly become a downward spiral of technical debt accumulation. It is mandatory to seize the given opportunity to redesign painful processes and corresponding IT systems. 

Spiderman thinking
Implementing RPA is not trivial

It can be tempting to think a problem is solved once temporarily not visible anymore. It can be even more tempting, if the project was successful (specifically if seen only from a pure business point of view), to make RPA the new core business component. For example, why not generating new reports from it, or even create new processes if flexible enough.

Rotten stilt house collapsed in water
Collapsed stilt house in Sattahip, entire neighborhood at risk

In the end, if not handled properly, RPA will only be a success for those implementing it and fixing the issues afterwards. Just like Levi Strauss became rich with his work pants during first gold rush... 

As a conclusion, RPA can be part of the answer if it is :

  1. Not perceived and used (only) as a cost reduction solution
  2. Not seen as one shot project but anticipates processes and IT evolution
  3. Part of a global transformation roadmap (business, IT, HR)
  4. Part of a clear strategy to define if it is outside of IT (and thus temporary) or part of IT roadmap (and thus have clear use cases and integration points)
  5. Of course not opposed to IT nor employees