The failure rate of Robotic Process Automation is 30–50%. Avoid falling for these mistakes now.
The following article was originally published in on January 5, 2018. To learn more about Robotic Process Automation, please visit our website.
What is Robotic Process Automation or RPA?
RPA refers to the use of software “robots” that mimic tasks usually performed by humans. These robots are especially helpful for automating rule-based processes that require interaction with multiple, disparate IT systems.
It is widely expected that in the Fourth Industrial Revolution, clerical work will be automated the way industrial robotics automated production in the shop floor.
With RPA being a relatively new concept (though it must be said that the underlying technologies themselves are fairly mature), we thought it might be timely to debunk some of the common myths or fallacies surrounding RPA.
Here we go…
1. There are no physical robots involved in RPA
The term Robotic Process Automation is, unfortunately, a misnomer. For no physical or industrial robots of any kind are involved. Instead, what you get are software robots (or just simply a software) automating digital tasks that are mundane and repetitive in nature.
The good news is that there is a shift away from Robotic Process Automation, and towards Intelligent Automation or Intelligent Process Automation. This is no doubt driven in part by our current obsession with all things related to Artificial Intelligence, but also the industry players’ desire to move up the value chain.
2. Robots will replace humans in the workforce
There is a palpable fear that technology in general, and robots in particular, will displace and replace humans from the workforce. And who can blame these employees.
Post the Great Financial Crisis of 2007–2009, job and wage growth in many economies have been tepid. Given this backdrop, the digital disruptions sowed by technological innovations certainly felt like a double whammy.
The truth however is much more nuanced. While casualties are certainly unavoidable, McKinsey estimates that less than 5% of all occupations can be automated entirely based on currently demonstrable technologies. A further 60% of jobs have at least 30% of constituent activities that can be automated.
What is actually happening is that robots are mostly doing the work that humans hated anyway. Or as the saying goes, “taking the robot out of human.”
And we have not yet considered that new jobs that will be created as the RPA industry grows and matures.
In short, we are unlike to see mass unemployment anytime soon.
3. Robots are 100% accurate
In a sense, the robots are really only as accurate as how the robot developer programs them to be.
This is because the RPA robots do not have any cognitive capabilities themselves. They will simply execute the set of instructions (to a T, it must be said) that they have been preconfigured with.
So if the robots have been wrongly configured or if some of the underlying assumptions concerning the robotic operations are erroneous, then errors will be inevitable.
4. RPA is (just) a glorified macro/script
While there are definitely similarities, a RPA software is inherently much more powerful than a simple macro script.
For example, the use of RPA provides the unique ability to integrate multiple, disparate IT systems at the User Interface (UI) level.
Furthermore, most of the enterprise RPA tools place a lot of emphasis on governance, which is not something you get with macros or scripts.
5. A RPA developer does not require programming skills
Many vendors are positioning RPA as a simple and easy to use software. One which any end users, subject matter experts or non-IT personnel with no programming skills can use to design and develop their own automated solutions.
Accordingly, it is not uncommon to come across tenders that specifically calls for a solution that caters for users with all levels of IT proficiency including users whom have little to no programming background.
Unfortunately, such approach often leads to the path of disillusion. For example, it has been widely cited that the failure rates of RPA initiatives are anywhere between 30 and 50 per cent.
Why is this so?
Many of the actual processes that needs to be automated are fairly complex with a number of interdependencies. Accordingly, the solution that is designed needs to be fairly robust with the ability to handle exceptions gracefully.
This is not something that any organizations can reasonably expect their normal users to accomplish. For example, a robot developer will often make use of selectors to interact with the various onscreen elements of an application. There is a fair bit of technicality involved in this process as you can see from the screen shot below.
Some astute readers might point out that many of the leading RPA software offers “record-and-play” functionalities. This basically a desktop recorder that watches the actions performed by a user and automatically generates the script.
However, the scripts so generated are typically of low quality and certainly not robust enough to account for various scenarios, and to handle possible exceptions.
That’s why Blue Prism, one of the leading RPA software providers, does not offer a desktop recorder. According to them, such an approach increases the Total Cost of Ownership over time, and reduces the longevity and resilience of the automation developed.
6. RPA is only about cost reduction
If you think so, then you are sadly missing the forest for the trees.
No doubt, one of the main benefits of RPA is cost reduction. A software robot cost anywhere between one-third the annual loaded wages of an offshore Full Time Equivalent (FTE) to one-ninth the annual loaded wages of an onshore FTE.
This is the main reason why automation is dramatically disrupting the labour arbitrage model of the Business Process Outsourcing (BPO) industry.
But the benefits of RPA extends way beyond cost savings. Other benefits include:
1. Increases speed to market as the software robots execute tasks instantaneously
2. Eliminates the risks of human errors
3. Ensures regulatory compliance
4. Improves productivity as the software “robots” operate 24/7 with minimal downtime
5. Flexibility to scale up and down the robotic operations depending on business requirements
6. Improves employee engagement by removing mundane, menial tasks from them
7. Transforms the operating paradigm from labor-intensive to technology-centric
Do review your RPA objectives and strategies if you have been concentrating on cost reductions all this while. You may not be extracting the maximum bang for your buck (or should I say robot).
7. RPA is expensive
As mentioned previously, the annual cost of a robot is lower than the loaded wage of a FTE. And that is not considering the fact that a robot works 24/7 with no lunch breaks, no annual leaves and so forth. Consider also that the fact that one minute of work for a robot is roughly equivalent to fifteen minutes of work for a human employee.
And RPA is definitely cheaper if you consider the other alternative — taking the traditional system integration approach where the total costs and project duration are so much higher.
For RPA, the first day is its worst day.
For SI, the first day is its best day.
Now, which one will you choose?
8. FTE stands for Full Time Equivalent, not Full Time Employee
Is this merely a matter of semantics? What is the significance of this distinction?
A lot, actually.
It determines fundamentally how you view your digital or virtual workforce.
A lot of organizations make the mistake of viewing the robot as a one-for-one (or better still, one-for-many) replacement of their existing employees.
Unfortunately, it does not work this way.
Commonly, we find that the average employee in a typical organization handles a multitude of tasks. Out of these tasks, some can usually be automated using RPA and related technologies.
However, there remains tasks which require the employee to handle. Typically, these are tasks that require human judgement, are not rule-based, or requires dealing with non-digitised or non-structured data.
Accordingly, post automation, a lot of companies find themselves unprepared to handle the delicate issue of retraining or redeploying the affected employees.
Don’t make this mistake. Involve your Human Resource and Corporate Communications departments, and be ready with your change management and communications plan.
9. Focusing too much on the technology
To ensure a successful RPA implementation for your organization (and we know how tough that can be), it is perhaps important to focus first on your people and processes.
Some of the issues you need to consider include:
· Have you obtained executive sponsorship of your RPA initiative?
· Do you have sufficient resources to implement RPA, be it in-source or outsource? This includes both hardware and software (people, licenses, etc.).
· How are you handling the communications to internal and external stakeholders?
· How are you managing the redeployment of employees (if any)?
· Which are the processes to be automated? Have you performed both the technical and commercial feasibility study?
· Have you established an appropriate methodology for your entire RPA implementation, covering various aspects such as ongoing maintenance, bots monitoring, change management and performance tracking (to name a few)?
· What is your business continuity plan?
· What is your governance model?
· How do you define and measure success?
10. Robots can automate everything
Most definitely not.
In fact, there are very strict guidelines on the type of processes that are suitable for automation using RPA.
Some of these criteria include:
· High transaction volumes
· Low exceptions
· Stable and well-defined processes
· Low system change
· Structured data and readable electronic inputs
Nevertheless, even if a process ticks all the above check boxes, it does not necessarily mean that automating through RPA will bring about real tangible benefits.
And that is because the underlying process itself may be inefficient and needs to be re-engineered first.
“The first rule of any technology used in a business is that automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency.” — Bill Gates
11. Robots can be left unattended 24x7
This will be the ideal end state, but can be difficult to realise in practice.
Typically, you will require a robot or process controller (a human) who is responsible for running, scheduling and monitoring the robots, handling exceptions, tracking robot performance, and all other production support activities.
Equally important is the need to establish a Business Continuity Plan. Because robots will and do break down, so you will need a contingency plan in place (which may not involve robots).
Which brings us back to the earlier point that robots will never eliminate humans from the workforce.
By clarifying some of the common misconceptions around RPA, we hope this helps to enhance your understanding of what RPA really is. And to empower you to deploy RPA within your own organisations to great(er) effect.
Do you know of any other myths about RPA? Do share in the comments below.