Stop and Think Before You Use RPA

With the growth in popularity of Robotic Process Automation (RPA), it can incredibly easy to rush into trying to automate absolutely everything. It is quite common initially for people to try and “over-automate”. Whatever the task at hand may be, people feel as though there should be an automated solution for it. However, that isn’t the reality. Some processes just aren’t suitable for automation. Also, some processes should be improved before they are automated. Automation isn’t a one-stop-shop solution, and it certainly isn’t a solution for an inefficient process.

Many companies are seeking improved efficiencies, process streamlining, and reduced errors through automation. The reality is that if you automate the wrong things, you will be fighting a losing battle. So, the question here is when is it not suitable to try and utilize RPA within the business? Let’s look at some potential scenarios:

The Process Shouldn’t Be Automated

Not all processes are suitable for automation. The standard rule of thumb is that anything which involves a lot of consistent manual input with repetition should be suitable as an RPA candidate. However, this is a general rule and isn’t a guaranteed result. Some processes just simply don’t suit being automated.

You should take a look at the process you are considering automating and ask yourself the following questions:

  1. How stable is the process? Does it go through changes, or diversions in the process regularly?
  2. How often does the process change?
  3. Is the process well documented?

As a general rule of thumb, you may be setting yourself up for more pain than it is worth if you try to automate something which is highly unpredictable or unstable. In saying that, it isn’t impossible to automate these tasks. However, my recommendation would be to always try to improve those processes so that they do become more stable if possible before you attempt to automate them.

A process that changes constantly generally isn’t suitable for automation. We’re getting to the point where AI is becoming a reality, and once that arrives, we will be able to handle tasks with a little less predictability. However, for the time being, true AI is likely going to require a significant custom cost implication, and will likely outweigh the benefit of the automation.

The Benefit Does Not Outweigh the Effort

This is incredibly common when someone first gets introduced to automation. The attitude can lean towards trying to automate absolutely everything. As discussed above, some tasks just aren’t suitable for automation, and the cost of automating can outweigh the benefit gained. There are also instances where the process just isn’t worth automating due to its simplicity, or manual requirements. An automated process can sit side-by-side with manual inputs, but sometimes the entire process is better as a wholly manual task rather than partially automated.

Having an effective KPI setupset up in order to decipher whether your automated processes are valuable or not is essential. However, all automated processes should be assessed for their likely effectiveness before they even enter production.

It is always worth measuring the potential ROI on an RPA project before you deploy it. It can save you significant headaches later on down the line when you realize the manual process was better, and you need to unravel your process to become a manual task again.

The Process Requires Improvement Rather Than Automation

It is easy to look at automation as the savior of all inefficiencies within a company. But the harsh reality is that sometimes the process just isn’t very good. This can come from years of using old processes and stitching new additions to that old process. This ultimately creates a poor process with a mishmash of other sub-processes bolted onto it over time.

Simply put, the best outcome can sometimes just be taking that process to the drawing board, and figuring out how to improve it. Sometimes there are steps that you no longer need. Sometimes an aspect of the process could be improved to make the whole process more efficient.

However, once a process has been improved, then perhaps you could look at options for automation. As I have already mentioned, trying to automate a task that is poorly set up, inefficient, or requires improvement can end up costing you more than the benefit of the automated outcome.

The Process Isn’t Valuable

Sometimes you just need to take a hard look at what produces value in a company. So many times I have seen processes executed which are completed by someone purely because “that’s how they have always done it”. Once you strip back the fact the process has always been there, and really look at the effectiveness and value the process brings. Often you realize that the process isn’t even needed.

Ultimately, removing a process that isn’t providing value will provide more efficient benefits to staff than implementing an automated process in its place.


I think the main conclusion you can draw from this article is that understanding your processes is the key. Constantly looking at processes and wondering whether what you are doing is the best way to do it. Just trying to chuck automation at a problem won’t necessarily solve it, it will just make a bad process an automated bad process, to begin with, An efficient process to begin with will be far more suited to an RPA implementation project than an inefficient one.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top