Robotic process automation (RPA) means automation of any process that is robotic or repetitive in nature. This article aims to clear any doubts about RPA and help you decide whether or not RPA is a good option for your business.
Has it ever happened to you that you were copy-pasting some form of data from one place to another, or even worse, manually writing everything. And after some time, when you rechecked, you realised that you made an error earlier and had to redo everything? Or you are an employer, and when you see your employees busy wasting their time with such mundane tasks, you can’t help but question if this time could be used for something better. We can agree that all such tasks are no less of an obstacle in productivity than slow Internet that makes you halt your work.
Or maybe all the things mentioned before made you nostalgic, but you are glad that you no longer have to deal with them, thanks to robotic process automation (RPA). According to US-based RPA provider Automation Anywhere, seventy per cent of firms feel that at least half of their workforce will be digital workers in a post-Covid-19 world. The survey that included nearly twenty countries had over 5,000 senior executives from mid-size and large organisations across India, the Middle East, and emerging markets in Africa and Southeast Asia.
But is it for everyone? Is RPA really worth the hype? If I invest in RPA, what should I keep in mind? Or first of all, should I invest at all? This article aims to clear any doubts about RPA and help you decide whether or not RPA is a good option for your business.
What is RPA?
In simple terms, robotic process automation (RPA) means automation of any process that is robotic or repetitive in nature. The automation involves the configuration of computer software or bots to mimic ‘most’ of the actions usually performed by a human to interact with disparate digital systems while executing a business process. The aim is to leave the time-consuming and mundane tasks to RPA so that the employees can focus on tasks that require more strategic and innovative thinking and meet the demands of rapidly evolving markets.
Sirisha, head of industrial engineering and new product launch, Continental Automotive India, says, “Covid-19 pandemic acted as a catalyst for digital transformation in the industry globally, including India. Automation eliminates many of the time-consuming steps involved in manual processes. It can reduce unnecessary costs and improve the quality of outcomes as well.”
But what are these tasks? If your work involves logging into different applications like e-mails, connecting to system APIs, filling forms, and simply copying and pasting data from one place to another, then these definitely meet the conditions of RPA. There is a video that explains how RPA is different from traditional automation at this link.
Types of RPA
There are numerous tools available in the market that automate simple processes to the ones that can be used to make complex decisions. RPA can be broadly classified into attended and unattended automation.
Imagine that you call the customer representative of a business for a query. What would you like—it takes a long time for the executive to manually find your details and make the necessary update, or it gets done in seconds?
This is the purpose of attended bots—saving time for performing repetitive tasks in real time, which does not follow a specific pattern. It can be on demand where a human involves bots at defined intervals for specific functions or in tandem, where the robot supports the human’s work process in the background. These bots run on computers to aid the employees in performing the tasks assigned to them. Invoking these passive bots has to happen either by the user’s action, or a window can be defined since the exact points of the launch are hard to detect. Control is then returned to the user by the bot to carry out the next set of activities.
By using attended bots, the executive comes up with the client details easily, but when your job is to enter data of several different new clients on various applications daily, it surely gets boring. For this, use unattended bots.
Unattended bots are best for the repetitive structured tasks that can be fully mapped and hence require no user involvement. In case of partially unattended, the human needs to define the parameters for an unattended process to work. This reduces the time taken by users in performing complex tasks that require dealing with high-volume data.
These can be self-triggered and even run on back-end servers 24/7 based on the occurrence of an event such as a predecessor task, a priority-based time slot, or even data input in a field in real time. One bot can even launch another. The best thing? There is no delay in task completion before deadlines.
The goal for businesses, especially large organisations, is to make support and back-end processes more efficient by using the best or a hybrid of both unattended and attended automation. So, basically an attended robot uses the human’s computer to start an unattended background process.
Want to know what else is an example of RPA? Try opening a few websites of different companies you wish to contact, and you will see a chatbot on the right-hand side to answer your questions in most of them. Most sites we visit these days have chatbots or virtual assistants that use natural language processing to analyse data in real time and interact with users. In certain instances, the data is given to the human professional for a better understanding of customer requirements.
RPA itself cannot learn from experience or interpret meaning like AI. As people are taking more interest in the integration of artificial intelligence (AI) with RPA for intelligent automation, performing cognitive tasks via machine learning (ML) that need handling of complex unstructured data (not organised in a database) will become possible.
Manish Bharti, president – India and South Asia, UiPath, says, “By combining attended and unattended automation, chatbots provide a rich user interface, by being able to converse with people and other robots in the workflow. Early this year, we partnered with Druid, and integrated their no-code, AI-powered chatbot into the UiPath Platform.”
Use cases of process automation in customer support services
Be it B2B or B2C, customers are the priority of a business. Recent studies show that the use of bots for customer service is going to further increase in the coming years since most processes like payment processing, email communication, and processing forms can be done with RPA. Take the example of system queries. Since any organisation will have several applications that are usually not integrated, manually compiling data would be quite tedious. RPA can collect data from all applications, including databases from servers, and provide the user with consolidated data.
It is based on AI and is a more advanced version of an auto-attendant system. It is great for handling automated tasks that involve numbers and simple yes/no answers. Voicebot with natural language understanding (NLU) based voice channel for communication works by converting audio to text format. For example, Dialogflow CX, launched by Google, is used for building text and voice-based AI agents for mobile apps, websites, and Google Assistant actions. A user can speak to the virtual agent naturally, and it will process the information and respond appropriately. Another example is the AI-powered voicebot TIA launched by Tata Capital on WhatsApp.
The AI-based human-like bot that pops up on the website to start an online chat is becoming a common feature on most business websites. The chatbot can be integrated on websites, messaging applications, mobile apps, and kiosks. These bots play a vital role in the simplification of the interaction of humans with computers and customer interactions with the business, providing engaged self-service.
As per some reports, sixty per cent of online customers do not like to wait for more than sixty seconds for a response to their query. The use of chatbots drastically reduces customers’ wait times and quicker resolution of their queries, thus improving customer experiences. Chatbots with natural language processing or NLP capabilities are used to handle customer complaints through faster responses, thus elevating customer satisfaction.
For example, the AI-based COTA (customer obsessed ticket assistant) system built by Uber helps customer support representatives improve their speed and accuracy, resulting in improved customer experience. It has been found that with the AI system, ticket routing efficiency is increased by ten per cent.
Some big companies such as Google, Microsoft, Facebook, and others are all actively engaged in the race to build chatbots that can respond to customer queries.
These include biometrics, face, and voice recognition. Biometrics refers to body measurements, authentication, identification, and access control. Facial recognition identifies and verifies an individual by comparing facial features from a digital image or video to a database. Voice recognition biometrics digitises words and encodes them with data such as pitch, cadence, and tone, and then forms a unique voiceprint related to an individual.
AI-based biometrics helps customer service agents to recognise customers and greet them in a personal manner. The customer support team can use biometrics to quickly authenticate customers while minimising the risk of fraud.
For example, tools like ScopeAI work with customer support teams to analyse the voice of customer (VOC) data across multiple channels.
Most companies now have access to huge amounts of data about their customers. When a customer clicks, views, and purchases a product from the website, these data are translated into predictions that deliver value-added personalisation and recommendations to targeted consumers. That is, AI-based predictive technology can identify patterns that indicate a customer’s intent based on Web activity or text, and route the call or chat to the appropriate agent.
AI has been used in the e-commerce industry to improve product recommendations based on personal customer needs, buying behaviour, and user profile. For example, Netflix’s AI-powered customer service algorithm uses data such as demographics, viewing history, and personal preferences to predict what the user would like to watch next.
This technology is used to predict technical and maintenance issues before major problems are developed. Some examples of AI predictive maintenance in use include United Parcel Service and Cisco Systems. It is reported that United Parcel Service has saved millions of dollars by implementing an AI predictive maintenance solution that reduces delivery truck breakdowns. In Cisco, AI predictive maintenance is used to optimise network performance and troubleshoot issues faster.
Business processes beyond customer support services suitable for RPA
So, are these business processes just specific to customer service? Absolutely not. Manish says, “Critical business processes such as contracting, compliance, customer service, salary payments are all tasks that are being automated. However, business processes such as finance and accounting—accounts payable, claims processing, data entry and reconciliation, document processing, and warehouse management—are a few areas that have witnessed the maximum deployment.” A few examples of RPA usage in different fields beyond customer services are discussed next.
Banking and finance
It is one of the biggest sectors for RPA as myriads of outdated processes and systems with varying levels of complexity can be automated. It becomes simple to integrate data from different channels consisting of documents in different formats (docs, pdfs, etc) with existing accounting systems. Common processes such as generating bank statements, KYC, loan processing, and invoice processing become much faster and more accurate.
This sector mostly involves variable activities that lead to a complexity to define RPA, sometimes even requiring AI. Tasks that can be streamlined include business intelligence reporting, digital ad positioning, basic sales procedures, and more.
Be it talent search, hiring and onboarding, employment history authentication, or absence of management, a lot of time is consumed in going through the same time-sensitive processes over and over again. Understandably, it becomes more of a pain point as the number of employees increases. Although new, a growing number of companies are deploying RPA tools in the HR department due to their quick implementation.
This is another sector where most of the processes like tracking shipments, shipping products, updating orders, among others can be automated. The bots can read and consolidate the required information easily. Hence, the bottlenecks related to returns and other complaints can be readily resolved.
Other industries like insurance and manufacturing can also benefit from RPA. Talking about the implementation of RPA in manufacturing, Sirisha says, “The advent of robotic process automation (RPA) has always been a key milestone for the manufacturing industry. RPA allows manufacturers to streamline operations and reduce managerial processes’ complexities, thus creating a more agile system. As a technology company, we have always believed in technology’s power—be it in the products we manufacture or in-house processes. We have implemented various Industry 4.0 practices at our plants, including a certain amount of automation across multiple levels. So, transformation to RPA was a natural step for us.”
Also, implementing automation based on departments instead of processes leads to duplication. For example, if both the HR and legal departments need the same list, there should be a single logic implementation.
Should you invest in RPA?
The economic crisis from Covid-19 has caused a sudden surge in RPA as businesses are looking for a cost-effective alternative to fill the workforce gap due to the shortage of employees. But as a senior decision-maker, the last thing you would want is losing money.
If there is a lack of data or the process is already obsolete, RPA cannot be a beneficial option. Successful automation is not just about having the latest tech tools but about understanding the current technology of your business and strategising beforehand.
It’s not just large organisations but also small to medium-sized businesses (SMBs) and small and mid-sized enterprises (SMEs) that are looking for new ways to get the work done efficiently. Although there would be more users doing the same work in large organisations, it is difficult for SMBs to retain talented employees.
Manish says, “Most businesses face pressures of new technologies, growth, customer expectations, cost pressure, technology and process optimisation, outsourcing, market saturation, and compliance and regulation. Yet, their core priorities stay the same, to increase revenue and lower costs, improve customer satisfaction, increase employee engagement, and reduce compliance risk. In a large organisation’s automation journey, there would be a full-fledged team required to take care of both the infrastructure as well as the application. However, investing in these resources and infrastructure can be a challenge for an SME that may not have the required capital or number of employees. Small businesses have the advantage of being able to situate themselves better due to organisational efficiency, process familiarity, and less chances for error. In the case of automating an SME, a cloud implementation comes to the rescue.”
Contrary to popular belief that RPAs are only for large organisations, the RPA market is evolving to become more affordable for SMBs and SMEs who are focusing on it to stay ahead of their competitors. According to a report from Grand View Research, the large enterprise segment accounted for more than 67 per cent share of the global RPA revenue in 2019, while BFSI led the market in terms of application with more than 29 per cent share of the global revenue. But over the forecast period from 2020-2027, SMEs are projected to witness the highest CAGR, with the pharma and healthcare sector on the forefront application-wise.
Sirisha says, “In India, most, if not all, automakers are using RPA in some form or the other. The influence of RPA would increase even more in the near future. Thus, RPA can be a vital tool for SMEs to increase productivity and save time. However, we need to consider the budgetary constraints here a little bit as the transformation to RPA can be very capital and effort-intensive in the first few steps. SMEs need to evaluate where they have a majority of redundant tasks and start from there. Steadily they can add more tools as they move ahead.”
However, there may not be a great demand for automation in very small companies that just started. By filtering the most inefficient processes, you can quickly identify how to best automate.
Okay, you think you should invest. What next? Ensure that RPA can address the gaps in the current system adequately and bring benefits to the organisation as a whole. The decision-makers can work with the employees to understand how a process works and the benefits RPA can bring.
Generally, time-consuming business processes that have well-defined rule-based steps and involve large volumes of data are suitable candidates for RPA implementation.
After finalising the processes that need to be automated, assessing ROI for individual processes is essential to quantify the benefits. The pace of change may vary from a few weeks to some months according to the business needs and scope of automation.
Finally, choose appropriate RPA toolsets and test them before implementing them into a real working environment. The bots can be run in manual mode initially so that errors can be noted. First, get started with the basics, and then gradually increase RPA deployment. Sirisha says, “Compatibility for integration with existing machinery and systems need to be verified before the investment plan for RPA. The other aspect to consider before RPA planning would be the constraint of standardisation for horizontal deployment across multiple locations of the same organisation.”
Here is a video that talks about how companies begin with RPA deployment.
The numerous advantages of RPA are the reasons you would want to go for it.
- Unlike humans, RPA can constantly work 24/7 and never make mistakes. It relieves the employees from non-value-adding tasks, increasing the speed and productivity of work. Moreover, even a non-technical person can be trained to automate RPA tools.
- Bots are reliable when it comes to complying with existing regulations and standards, and their operation is continuously monitored.
- Unlike traditional automation, RPA uses software bots that can be easily programmed to understand user actions.
- It requires little integration, that is, it can be installed on a desktop or cloud environment, making it suitable for remote working. It is scalable as multiple operations can run simultaneously, and additional bots can be added depending on the workload.
- Finally, companies deploying RPA can witness a positive ROI within a year or two.
When talking about SMBs and SMEs, there might be a tradeoff between the feature and budget.
Although there are tons of vendors in the market, the tool that fits your budget might not necessarily mean that it has all the features that you need for successful automation. This is especially true at a time like this where it is difficult to raise funding. In such a case, deploying the tool would rather cause new problems than solving the existing ones.
The time taken for complete implementation can be quite different in theory and practice based on the process choices made. Additionally, their maintenance requires extra expenditure as well as time, which can be a daunting process.
The majority of the solutions available in the market have limitations when it comes to adapting to a changing business process, making it difficult to find a scalable and customised solution for your business.
While AI integration is gradually increasing, most of the tools cannot still take on advanced cognitive tasks that need learning from unstructured elements. Such tasks often lead to a complex RPA approach that is difficult to define and implement.
How to assess ROI
Assessing ROI depends on whether you want to calculate the ROI for a single process or a complete programme. Then build a business case for your project, which compares benefits against the costs incurred in the upcoming weeks or months. Quantitatively speaking, the most basic formula for ROI would be the difference between the cost of automation and the cost of labour, taking into account the number of hours spent in manual work.
First, let us talk about savings.
The major benefits are the time saved and the elimination of errors. The time taken by employees to manually work on processes such as data transfer is saved by automation, and they can instead focus on their core jobs. The FTE (full-time equivalent), which is the number of working hours of an employee in a fixed period of time, can be adjusted to incorporate the added productivity from the returned work hours. RPA also does away with factors such as the use of IT resources and training costs. When automation reduces the tiring work of employees, it increases work satisfaction, which means reduced staff turnover and better productivity.
In case an employee is on leave or there is a request overload, it leads to downtime when ideally the service should have been available. The improvement due to the shortening of queues, service availability, and customer satisfaction can be estimated and added for ROI calculation.
Doing away with manual errors means a reduced need for cross-checking and reducing the number of redos for a task. Depending on the error rate of a process, you can add that into the estimation too.
On the flip side, RPA doesn’t come for free. So, consider the costs involved during and after deployment.
For the cost of RPA, first, consider the license cost of the tool. If the process is dynamic, more bots will be needed, and hence the cost will be higher.
Then consider the cost of infrastructure required for deployment after consulting the IT department—servers, test environment, and so on.
Add the potential cost of development, taking into consideration as many factors as possible over time. Based on the complexity of an individual process (low, medium, or high), the average development time and expenditure will differ. Whether a company decides to partially or fully outsource the development—training or hiring experts, running the automation tests, etc—will change the cost substantially. With experience over time, the teams will perform better than they did initially.
Finally, the maintenance and monitoring costs come into play when there are issues in the additional infrastructure, changes in business process needs, exceptions encountered, or as a result of any other unexpected outcome. The subscription-based cost of an RPA tool includes maintenance issues related to the tool itself, but these problems are not covered in it.
If this sounds like a lot of work, there are numerous ROI calculators available online to help you with the estimation. Depending on the delivery model, the ROI may vary, but the time period is always shorter as compared to other forms of IT automation. Faster ROI, sometimes even a few weeks for a single process, is one of the main reasons why companies are willing to go for RPA. According to McKinsey, they noticed that the ROI varied between 30 and 200 per cent in the first year across sixteen of their case studies.
Numerous vendors offer free trials and even cloud RPA as a service for greater benefit.
Choosing the right tool
There is no single right tool for a particular type of business. It just depends on your needs, the ability of the tool to integrate with existing systems, and the costs incurred. As mentioned earlier, the total cost includes the setup cost, license fees, and the maintenance cost for the tool.
It is good to check if your preferred vendor has successfully deployed bots in a company in the past that had similar pain points and was similar in terms of size. Also, read the terms and conditions for maintenance and support and ensure that your requirements are met.
The chosen tool must be platform-independent and able to support different types of applications. It should have a user-friendly interface so that employees can get used to it easily, and there are no hassles in learning. RPA tools can be prone to attacks if proper security measures are not in place. Scalability is important for automation to be efficient at a large scale.
“I remember an industry senior telling me how the RPA market size will grow in the coming years during my college days. While many fear that it will replace humans, experts believe that RPA is more likely to alter current job functions and even lead to new job creation. Automation with its present capabilities cannot be a substitute for problem solving and interactions that only humans can perform. Although the fact that employees will need to upskill and reskill to fit in the dynamic job roles cannot be denied,” says Sirisha.
Sirisha further says, “Whenever we speak about new technology, it is critical to ensure the employees are upskilled adequately for it, and the change is celebrated in organisational culture. Employees handling processes starting from the warehouse’s material movement until the technology-intensive process like testing need to be upskilled and gain know-how about using the automated processes, which helps save time. For example, engineers handling tasks like total productive maintenance and analysis understood the new infrastructure and adapted themselves to new maintenance and defect analysis workflow.”
Manish says, “With the onset of the pandemic, we have also seen a large uptake from governments as well. For example, EY India with UiPath developed an RPA solution for the Municipal Corporation of Greater Mumbai to generate bulk e-passes for service aggregators supplying essential supplies and services to residents of Mumbai during the lockdown. This helped relieve their manual team from investing tremendous efforts in mundane, time-consuming tasks.”
While employees’ workstyle will definitely change with automation, on the technical front, RPA, just like any other technology, is prone to cyberattacks. This can lead to a data breach. But if proper measures are implemented, it can rather serve as an asset.
“Technology is a boon when used in the right manner. It can be and should be used to make our lives easier and safer. Cybersecurity is becoming predominantly crucial in the production environment of various processes like the end of line stations, test stations, etc, where data security is essential. At the manufacturing unit, the level of standardising cybersecurity systems is key to successful deployment,” adds Sirisha.
RPA architecture needs to have multiple protection layers, so that the RPA tools won’t open a suspicious link. This reduces the chances of human-centered attacks as the access to authorised information is with RPA. Proper monitoring and alerts can keep the users updated in case there is an intrusion.
Yes, a combination of RPA, AI, BPM tools is worthy for automation. As one thing is clear. The future is automated. We all have to adapt and prepare for it in all parts of our society.