Enterprise Resource Planning Blogs by Members
Gain new perspectives and knowledge about enterprise resource planning in blog posts from community members. Share your own comments and ERP insights today!
Showing results for 
Search instead for 
Did you mean: 
It's an exciting time to be living in for tech enthusiasts. Never before had so many new technologies disrupted businesses and consumers alike. For enterprises, it presents a new set of challenges - the decision to choose which technologies to adopt and how much to invest in each. What makes the decision even more complex is the interplay and dependencies of these disruptive technologies with each other which makes it difficult to do a cost-benefit analysis of any of these technologies when viewed in isolation.

One of the hottest new development we are seeing is in the field of automation applied to business processes - Robotic Process Automation. Another technology that is poised to transform businesses is machine learning. I am not delving into the definitions of these since the internet is awash with literature explaining these in detail. (although if you are not familiar with these terminologies I would urge you to get a fair understanding of it before coming back to this article)

The primary focus of this article is to study the interplay of these relatively new technologies with two traditional enterprise software tools - ERP and BPM - and help businesses decide the right RPA tool in this context.

While RPA has already seen significant interest and adoption in transaction intensive processes such as back-office operations in banking and insurance, its adoption in processes driven by ERP systems has remained largely muted, especially considering most large organizations are run on ERP today. But this may be about to change. As increasing number of RPA vendors design solutions tailor-made for ERP software with special focus on SAP, the scope of applicability has broadened and the cost of ownership has reduced. In fact, the TCO of RPA and the time taken for operational and financial value realization is much lesser than ERP (although, admittedly, the scope is that much lesser too).


An organisation considering the use of RPA for ERP driven processes will be well advised to use it in conjunction with a BPM tool with a flexibility of extending it with machine learning.

The interaction between the systems can be explained as below:

  1. The ERP system forms the backbone of the IT landscape by acting as a single source of information for the enterprise.

  2. RPA (Robotic Process Automatic) - The RPA tool itself acts like a human interface and interacts directly with the ERP system just as a human would – by entering transactions, executing programs and extracting reports among other things.

  3. BPM tool should capture the detailed business processes, including processes designed for ERP, and define the interfaces between ERP system, its satellite systems and other standalone applications. BPM is especially useful for RPA deployment because automation should be decided keeping in mind business process and not user. A correctly mapped BPM model, therefore, makes it easier to select processes for automation using RPA. In many ways, RPA is an extended arm of BPMS.

  4. Machine Learning - Some part of machine learning is inherently part of RPA since RPA is 'trained' to act like human interface based on historical data. However, with the kind of advancement we are seeing in machine learning, it becomes imperative that RPA should come built in with flexibility to update its machine learning capabilities. Machine learning, therefore, acts like a feeder system for RPA which enhances the functionality of RPA and will improve its performance the more it is used. One of the reasons for the lack of interest in RPA among ERP customers is that most of the processes that run on ERP could only be semi-automated until recently. This is changing drastically with progress made in machine learning and will be a decisive factor for the success of RPA being used with ERP.


Other pointers that must be considered before going for selection, or even decision, of deploying RPA for ERP-based processes are briefly stated below:

  1. It is important to clearly define the goals of automation. For instance, for one organization, periodic closing may be of primary importance (time factor) but for another, the focus would be reduction in the number of FTEs (full time employees). Evaluate the solution offerings from different vendors based on this assessment.

  2. Evaluate the promised benefits vis-a-vis robotic process benchmarks. As mentioned earlier, the benefits derived from RPA are more measurable than, say, an ERP or BPM software, so comparing against standard industry benchmarks will help figure the approximate financial and operational benefit.

  3. Level of automation possible - This will largely depend upon the kind of processes typical with your industry. For instance, in a manufacturing unit, a lot of time is consumed in entering production and maintenance data into ERP system although the data may be readily available in another system. So, an RPA vendor offering the possibility to read from one system and feed into the ERP should be preferred even if it requires a little manual intervention.

  4. Security - Your ERP holds the most business critical data of your enterprise. It goes without saying that a water-tight assessment of the security features must be done before zeroing in on your choice of vendor. The bare minimum should be a vendor that provides complete audit logs, relies on the industry standard Transport Layer Security (TSL) 1.2 protocol, and employs latest encryption technology.


Robotic Process Automation is the next big wave of cost and process optimization after BPO. As mentioned earlier, the number of RPA providers designing solutions specific for ERP system is only increasing. An organization that wants to unlock the full potential of their ERP deployment should proactively consider using RPA for automating their ERP functions. The question that the CIO should, therefore, be asking now must revolve around 'Which?' rather than 'Should?'
Labels in this area