Is process “X” suitable for automation?
What are the criteria or factors that can be used for the decision?
When you start considering automation, the first challenge is to identify candidate processes that are suitable for automation. The challenge is knowing which processes are the right ones to automate.
One approach, could be to discover the processes that have been automated by similar businesses or competitors. This might be difficult to achieve as businesses do not publish such information.
A second approach would be to look at the case studies published by automation software vendors. These contain some broad indications of the processes which have been automated.
Both these approaches can provide some guidance, but they will raise the question of whether the processes in your business are actually the same as in other companies.
A third approach could be to engage a process consultant to review and document all your processes, asking them to recommend where automation would be appropriate.
Any investment in reviewing and documenting processes will deliver its own benefits. However, it is rare for a “process” expert to also be an “automation” expert. They will rely on experience, checklists and other criteria to identify the processes.
With access to checklists and guidance on the factors, would you be able to select suitable processes? Yes, of course you would. This could be the fourth approach.
A fifth approach would be to use data to provide evidence of processing activity that is suitable for automation. There are some sophisticated “process mining” tools available that are used to deliver that information. Once purchased and installed, they monitor the activity of many users in an environment and look for patterns in the data.
This article aims to provide some knowledge to perform your own candidate process analysis in respect of selection for automation using RPA (Robotic Process Automation).
Overall selection considerations
There is no one factor that determines suitability.
Generally, automation will only be implemented where it offers a cost-effective solution, but there can be exceptions. For example, one manager required a process to be automated as the current activity meant he was working in the office during the early evening. For him, the automation enabled him to get the work done while he was getting home for precious family time.
The selection factors have been grouped into:
There is no implied order of importance and no weighting between the groups.
Quantitative selection factors
- Does the process only use digital data as input and only produce digital data as output?
- Are all the decisions based on rules/criteria using data?
- Do any decisions require subjectivity/emotion?
- Are more than two computer applications involved?
- How many computer screens/windows could be seen in the process?
- How many data files are used in the process?
- How many data items are involved in the process?
- How many staff perform the process?
- How many times a day/month Is the process performed?
- Could the process be undertaken overnight or at weekends?
- If a process is currently undertaken overnight or weekends, what is the constraint driving this time?
- Is the process currently outsourced?
- Do instances of the process ever get paused?
- Is it important who completes the process?
- Are there management reports on the state of the process?
- Are the processes subject to audit reviews and external compliance checks?
- Are legacy applications used in the process?
- How many variations are there in the process?
- Does the process involve any data matching or comparisons?
- Is there more than one trigger for the process?
- Could the process be divided into discrete tasks which could be executed separately?
- Is the process documented?
- Is the cost of errors in the processing high?
- Is it tedious to perform the process?
- Are the applications which are used by the process highly available, stable and not subject to frequent change?
- Could the tasks be easily given to a new employee?
- Is there time pressure to complete the process?
- Is any of the data being processed sensitive?
- Can the process be reversed?
Using the selection factors
In this article no scores or scales have been used, so that the factors can be widely applied.
It is generally agreed that automation is about “taking the robotic tasks out of the human activity”.
In my experience, one of the best ways to use the selection factors is to consider them in the context of several candidate processes. This provides a comparison.
Once suitable processes have been identified by the selection factors, then consideration can be given to the costs of implementing an automation solution and the benefits that would be delivered. The selection factors may help identify where particular benefits will be realised.
A comparison of the financial case for automation can be conducted and this is likely to set the priority for an automation programme of work.
As experience is gained in using the selection factors, the question about when to use automation (i.e. RPA) can be answered as a rule of thumb:
When the process to be automated is repetitive, stable, well defined, of reasonable volume, uses multiple applications and the users would benefit from a quick implementation.