Robotic Process Automation, usually abbreviated as RPA, is growing in popularity with businesses and knowing what RPA is and what it can do is important. Career options in RPA make it a lucrative area for people to consider with potentially a lower learning curve than other area in information technology.
What is RPA? RPA stands for Robotic Process Automation and uses software ‘robots’ known as ‘bots’ to mimic human actions involved in carrying out manual tasks. This allows manual tasks to be automated by using these RPA bots, allowing the tasks to be repeated without any human interaction, thereby reducing the chances of any errors occurring. Monotonous time consuming but repeatable tasks are good candidates for RPA.
The RPA bots perform the tasks on a computer, using the same interface a human worker would, clicking as necessary, typing, opening applications, and using keyboard shortcuts in much the same way as a human operator would.
The intention behind RPA is to remove people from onerous computing operations, and it is predominantly used to automate business processes and tasks, resulting in reductions in spending and giving businesses a competitive edge.
RPA is a transformational technology that is fast introducing automated digital labour to the workplace in many different levels and competencies and promises to be as seismic as anything from the industrial revolution.
RPA effectively bridges the gap between completely manual and fully automated systems and uses intelligent insights to select what is worked on or selected from data streams.
RPA has evolved quickly over the last few years[1], and has become an important part of everyday operations in business, but how did it become so important so quickly?
When was RPA started?
RPA of today[2] started from computerised automation, which became an important factor in mainframe computing in early in the 1970’s, and which required a variety of software applications to perform essential operations. The problem was that the computations were becoming increasingly complex, and file sizes were growing.
As computer were being put to increasingly large tasks, the data flow became massive and the need constant review for results, and humans were not the best means of doing that.
The answer to the problem came from automation of some systems, allowing data to be searched and simple operations to be performed automatically, releasing humans from mundane but necessary tasks.
By allowing the computer to carry out at least some tasks itself, seemed like a novel yet elegant approach; freeing up humans was, after all, what computers had been originally envisioned to do.
What are Advantages of Computer Automation?
The five major advantages of computer automation were seen as:
- Reduced operational costs. By releasing staff to carry out other tasks, the computer centre would realise a greater overall efficiency, thereby reducing company costs.
- Increased productivity. Since a computer is capable of carrying out a mundane task many times faster than a human, it follows that there will be a boost to productivity. Particularly as released staff can work on other areas, allowing at least two tasks to be completed in tandem. The more tasks a computer can handle silently, the more complex tasks the human operators can tackle.
- Allowing high availability of resources. One of the first tasks to be included in the computer automation banner was simple save and recovery operations which ensured that data was backed up to the save medium regularly. This easy but necessary task allowed IT staff to have data backed up without having to even think about it, ready to be restored should the operating system crash, which it did, often.
- Increased reliability. Before computing systems were even an operating system (such as Windows) they were high level languages written running in a fundamental directory operating system (DOS). By the late 1960’s languages such as Fortran, BASIC, and COBOL were in massed use, and the almost universally-employed C language appeared in 1972. As computing stated to tackle larger projects, databases started to appear, and the SQL query language started to come into common usage. These high-level languages help programmers automate increasing aspects of everyday computing.
- Optimised performance. When carrying out complex programs, a system may need as much programmer attention as possible, and by automating certain parts of the background processes, the IT staff could spend their time carrying out important tasks rather than the mundane.
As the use of computing grew in business, and extended in even small companies, sub-routines that help complete low-level tasks became increasingly important and grew into the basis of RPA.
Business Process Management
By the 1990’s simple process automation had grown into a much more defined system called Business Process Management, which was now being routinely used to engage many different lower level processes.
The use of operating systems such as Windows 3.0 and Mac OS on microcomputers further automated certain routine functions and allowed the user to get on with higher level operations without having to worry about completing certain set processes.
Business Process Management liberated computer users from simple tasks and sped up the use of computers. It was by this point that personal computing and small business users were starting to really feel the benefit of computers and start to make them perform in a way that was useful for all users.
With Business Process Management being such a success, the opportunity to extend the processes used by the system was obvious. With computers becoming ever more powerful, and Central Processing Units (CPU’s) able to handle an increased number of operations, it was clear that the automation of sub-processes could be taken much further.
This led to a whole new industry built around the development of processes and investigations into exactly what was achievable in terms of automation.
RPA screen scraping
One of the first major tools in RPA was known as screen scraping and was defined as the process of collecting screen display data from one application and translating it so that another application can display it. Essentially and important to the growth of RPA. was the fact that screen scraper software could also work across platforms.
For example, screen scraper software is available to take the output from a legacy application running on an IBM mainframe and use it as input for an application running on a PC. This allowed the software to be used on multiple system simultaneously.
RPA grew from both computerised automation and business process management to become a truly automatic process and has become so good at carrying out routine tasks that some have questioned whether it isn’t actually the start of artificial intelligence (AI).
RPA Alignment with Artificial Intelligence.
At this point, it is necessary to review the connection between RPA and artificial intelligence and look at the cross-over between these two sectors, particularly as there is a rising use of the term Intelligent Process Automation, which implies that RPA and allied processes have a degree of intelligent operation.
Definitions surrounding these different types of learning machines are found in IEEE Standard 2755.1-2019 – Guide for Taxonomy for Intelligent Process Automation Product Features and Functionality.
This standard sets out to differentiate between the types of machine-driven applications and help consumers in assessing and selecting products that best fit their needs. In this way, the standard makes quite definite distinctions between AI and RPA. The standard defines RPA as:
“preconfigured software instance that uses business rules and predefined activity choreography to complete the autonomous execution of a combination of processes, activities, transactions, and tasks in one or more unrelated software systems to deliver a result or service with human exception management.”
While AI is described as:
“the combination of cognitive automation, machine learning (ML), reasoning, hypothesis generation and analysis, natural language processing and intentional algorithm mutation producing insights and analytics at or above human capability.”
Simply put, the standard considers the main difference between AI and RPA to be that RPA is a software robot that mimics human actions, whereas AI is concerned with the simulation of human intelligence by machines. That definition shows, quite neatly, that RPA and AI are quite definite and separate entities.
However, while RPA and AI are very different parts of the same overall ethos, there is a very significant connection in the use of cognitive aspects, so there is always a connection between these two facets.
When can RPA be used?
Obviously, RPA has its greatest use in fields where the business deals with vast amounts of data, and which accuracy is absolutely critical. This means that it finds use in sectors such as banking, finance, and insurance.
Financial companies are using RPA to optimise a wide range of back-office processes and fulfil time-consuming manual tasks that, until recently, required a great amount of costly human input. But there are other distinct advantages to using RPA in these kinds of environment too:
- Close to Zero percent error rate. Tasks undertaken and completed by RPA tools are phenomenally accurate and have a nearly zero percent error rate. This means banks experience fewer customer complaints and helper customer service departments deliver a seamless customer experience.
- Operational efficiency. Banks and other financial institutions process a very high volume of customer service requests on a daily basis. By using RPA, they can automate the manual and repetitive tasks that fill the daily operations of customer service functions. Furthermore, they can increase both volume and efficiency with checking order status, opening cases, record updates, and many more daily operations too. Contact agents can be more efficient as all information is automatically synchronized between systems since agents have all customer information to hand and, therefore, won’t inundate clients with questions and requests. This likely to manifest as increased customer satisfaction.
- Allow Focus on Customer Relations. As stated above, front end RPA solutions enable banks to expand their channels of communication and also to increase the confidence of customer-facing departments. This can take the form of either increased customer awareness and/or fulfilment of customer requests. For example, customers can request balance inquiries using other channels such as text and be supplied with the data via an RPA system. Such RPA systems automate the repetitive tasks agents deal with on a daily basis, saving them time and energy and allowing them to develop more meaningful relationships with customers.
The fact is that banks and insurance services deal with many thousands of customer queries on a daily basis and are expected to respond quickly. To be able to do that without a high level of data retrieval would not only be unreasonable but also practically impossible.
Banks use RPA technology to automate rule-based processes such as accessing and searching account information, customer application status, and balance information to respond to customers in real-time and reduce turnaround time to seconds. Because RPA bots are able to search millions of records for the pertinent information both quickly and accurately.
Some institutions[3] have demonstrated the use of RPA during the customer onboarding, the process of taking a new customer on experience, to automate the new account and deliver a positive feeling. RPA is seen as a credible means of speeding up what would otherwise be long-winded and tedious administration tasks, while reducing the possibility of mistakes being made.
Outside of customer services, there is increasing use of RPA in processes such as trade settlement procedures, which are processes where securities or interests in securities are delivered. Usually this against some form of payment of money, to fulfil the contractual obligations, arising under these securities trades.
Since a bank may experience a high volume of these, and mistakes cannot be made, this is an ideal opportunity for bots to perform this precision task. As far back as 2016, banks such as BNY Mellon[4] began to embrace the concept of RPA to carry out transaction such as settlement procedures, and soon started to feel the benefit in terms of closed deals and customer satisfaction.
Because of RPA, the company saw an 88% improvement in transaction processing times and account closure validations carried out across five different systems, leaving staff to tend to other parts of the business.
With regulation being a major part of banking and financial institutions, and a sample[5] of compliance officers found that around 73% believed RPA to be a positive inclusion in a company and will improve compliance testing and overall result as it becomes further entrenched in businesses.
Banks are increasingly using RPA tools to collect information about transactions and analyse the quality of such transactions against a set of specific validation rules. If a suspicious transaction appears, the robot instantly reaches out to the compliance department that can handle the issue. By flagging such problems quickly, RPA can help save financial institutions from reputation damage.
Is RPA cost effective?
Of course, there is always going to be a fundamental cost issue with RPA, and this splits down into two distinct areas; purchase cost and people. The cost of software may be initially fairly small, but as data grows, as company may need to invest in other automated equipment, such as automated tape librarians or automatic cartridge loaders, additional software, and messaging services and so on to allow the RPA full access to the data library.
The cost of people might be initially quite high, but this is likely to remain fairly constant rather than rising over time as with hardware and software. The main cost with people is the initial installation of RPA software and any associated hardware, and this may well be carried out by a specialist contractor.
Once the system is installed and commissioned, human intervention comes only from caretaker roles or troubleshooting any initial issues. Generally, interrogation of the system and the addition of new information into the database is usually the role of customer-facing personnel, and rarely takes the time of computer or IT specialists. This leaves the IT department in a position of overseeing and occasional maintenance.
RPA, being mostly software based, is envisioned as a self-supporting and requiring little human intervention during normal operation, so once purchased and installed the only ongoing costs arise form additions and upgrades.
These can be carried out either by in house IT or specialist contractors and represent only periodic expense. Because of this, RPA is a hugely cost-effective option, particularly when weighed against the saving in both time and money, in other modes of investigation, which may be personnel heavy.
RPA and Big Data.
Big Data has become an increasingly important issue in many sectors of business, particularly those that hold and sort huge amounts of business information. What actually constitutes Big Data is defined by the variables;
- Volume: Organizations collect data from a variety of sources, including business transactions, smart (that is, the Internet of Things) devices, industrial equipment, videos, social media and more. In the past, storing this volume of data would have been a problem, but the advent of cheaper storage on platforms like data lakes and Hadoop have eased that particular burden, and volume is now not an issue.
- Velocity: With the growth in the Internet of Things, data streams into businesses at an enormous speed and must be handled in a timely manner. With sensors, smart meters to RFID tags, are all driving large amounts of data in near real-time speeds.
- Variety: Data comes in structured and unstructured formats, from structured data in traditional SQL databases to unstructured non-SQL type data from text documents, emails, videos, audios, stock ticker data and financial transactions.
Because of the commercial interest in data mining, more organisations are storing, processing and extracting value from massed data of all forms. Systems that support large volumes of both structured and unstructured data will continue to rise as data becomes not only more available but retains an intrinsic value.
The market will demand platforms that help data custodians govern and secure big data while empowering end users to analyse that data and extract meaningful information from the mass of information. As these systems continue to operate inside of enterprise IT systems and standards, tools like RPA become essential to interrogate the huge amount of data available.
Some market commentators[6] have argued that RPA isn’t a simple plug and play option and requires a lot of specialist setting up to make it truly effective for data mining. However, once it has been correctly implemented, then it becomes a powerful tool for searching the mass of data that is usually contained in such databases.
In fact, usually, the data sets are so large that searching by any other means is impractical, if not impossible, for even low levels of enquiry, but if the system is dealing with many thousands of them, then an intelligent search system is essential.
Big Data is a growing field, and RPA is growing with it. The unwieldy size of many databases prohibits searches by humans, but is ideal for RPA technologies, and this will only become more so as the use and size of databased increase. This was a role that RPA was truly designed for.
Impact of RPA on employment
As already stated, software tools such as this are bound to have an impact on a company’s workforce, but from a positive point of view. RPA software robots, much like their mechanical cousins, can quite easily be seen as threats to human jobs.
The jobs at risk are the low skilled repetitive jobs, just like industrial robots replaced similar types of jobs in the manufacturing industry. Jobs which people hated to do can now be carried out by autonomous computer system, freeing operators up from tedious tasks.
RPA will usually effectively reduce a company’s headcount in a static situation, but the real situation is that many data centres are already suffering from a skills shortage in what is an increasingly complex and dynamic environment.
Many tasks can no longer be handled manually as their sheer volume and velocity increase, and data sets become impossibly large. RPA and AI in its various forms are already taking up the spare room and deliver an excellent experience that is far more accurate and faster than any human search can ever be.
Does RPA have future?
RPA is unlikely to remain at the level that it currently operates and will increasingly encroach on other areas of data review and selection and the customer experience, and in doing so, will show its real worth. As that happens, it will become increasingly prevalent in a growing number of industries, and that is likely to have an impact on IT personnel levels.
It is never a good idea to simply offload highly trained personnel like IT specialist, but also almost equally difficult to justify them in an increasingly automated system.
One possible answer is to modify the role of the IT specialist to make them more relevant to the changed situation. This can be achieved by giving your operations personnel new responsibilities and attendant titles, such as operations analyst, networking technician, or PC administrator. It keeps them in a position where they can be helpful to the running of the RPA system, without demeaning them.
As the old positions are no longer needed, the new technology gives rise to new and greater responsibilities and the opportunity to explore how the RPA tools can be best utilised. RPA-using companies need to realise this and maximise the talent they already have.
This means that, just because technology moves on, that people are not still required, though the roles that they actually pursue with an RPA-using company may have to be modified to suit.
How big is the RPA market?
A report[7] by Statista about the size of the RPA market, shows that, globally, the revenue growth in RPA will reach $10.4 billion by 2023, and will continue to grow exponentially from there. This will be driven by both the growth in data retention, and the need to search it, and the ease with which RPA technology can be implemented, so that it becomes viable for even smaller businesses.
Global management consultants McKinsey and Company recently issued a report[8] which predicts that processes such as robotic process automation, will have a potential global economic impact, that is, an effect on company business to the value of nearly $6.7 trillion by 2025.
The growth of RPA is accelerating, and it is likely to become one of the leading technological platforms and is expected to become the standard operating model for both positive business outcomes and performance.
Conclusions
Human error has been the cause of many fundamental mistakes with areas of high data flow and retention. By removing this element, the probability of creating operations that are not only more accurate but faster and more streamlined
Systems like Big Data are not only driven by tools such as RPA and would be unlikely to exist in any really useful form without such without them. Certainly, having masses of data without the tools to effectively mine them would be pointless
According to forecasts[9], the digital universe of data is likely to grow substantially over the next few years, and expand exponentially from there, with much of this information collected in large institutions in the financial sector. As data becomes amassed, RPA tools will need to grow and become smarter in themselves, to be able to cope with the demands placed on them.
Productivity is key for management, so with more reliable processes, that can be done repeatedly many more times over, are powerful arguments for the adoption of RPA solutions. However, achieving these benefits and operating them on a daily basis requires discipline to overcome the obstacles, particularly from the changed employment status.
As long as a business understands, anticipates, and ultimately balances these obstacles against the potential benefits of automation, they should not interrupt any future plans.
RPA is truly an example of modern-day Pandora’s Box. We have now released the notion of robotic process automation and have demonstrated its benefits, and those are powerful arguments for its continued use.
We are now likely to see it grow outside of large, data-driven companies and become embraced by any company that routinely searches data for customer information. RPA is now going to become a standard tool for companies of any size.
References:
[1] https://www.processexcellencenetwork.com/rpa-artificial-intelligence/articles/rpa-and-cognitive-automation-only-for-big-business
[2] https://nandan.info/history-of-robotic-process-automation-rpa/
[3] https://www.delaware.co.uk/en-gb/blogs/how-rpa-can-facilitate-onboarding-and-strengthen-company-values
[4] https://www.bnymellon.com/us/en/newsroom/news/press-releases/bny-mellons-automation-efforts-draw-industry-accolades.jsp
[5] https://www.accenture.com/_acnmedia/PDF-10/Accenture-Compliance-Risk-Study-2016.pdf
[6] http://tdwi.org/articles/2018/06/18/diq-all-how-robotics-process-automation-eases-data-management.aspx/worldwide-robotic-process-automation-market-size/
[7] https://www.statista.com/statistics/740440/worldwide-robotic-process-automation-market-size/
[8] https://www.mckinsey.com/business-functions/mckinsey-digital/our-insights/disruptive-technologies
[9] https://hostingtribunal.com/blog/big-data-stats/
