The term “robotic process automation” (RPA) was coined in 2012 by Blue Prism, the company that had pioneered the automation of office management serviced 17 years ago. However, at that time, the idea of RPA was not taken seriously and was considered as a utopia due to the extremely high cost of implementation and a very high entry barrier.
But everything changed with the advent of artificial intelligence (AI) technologies. It was them that fundamentally changed how businesses perceive RPA and its overall value.
This article will explore the essential steps you need to make before investing a single dollar in RPA implementation.
RPA Today
According to Grand View Research, it is expected that RPA implementation will cost companies more than $3 billion by 2025.
Modern RPA systems can fully or partially automate the work that previously had to be done manually. Let’s take a look at a bank with over 200 employees, for instance. For 60% of their time, bank employees enter customer data into the internal systems and transfer data from one system to another, print out tons of paper documents such as receipts, bills, and applications. Good news is that all this routine can be automated once and for all and any business can win from this automation.
How RPA helps facilitate different functions and operations
So how can a robot assist with driving retail, fintech, or insurtech businesses forward?
- Open and close web client and email applications and systems (including attachments);
- Move data across applications (ERP, CRM, etc.);
- Create and move files and folders;
- Follow links and simulate button clicks;
- Perform automatic transitions between applications;
- Fill in and copy forms;
- Upload data from external sources to the program interface and directly to the database;
- Compare and check data, perform complex mathematical calculations;
- Create complex logic with conditions and cycles;
- Recognize a myriad of text formats (e.g.,PDF / DOC / XLS);
- Perform a multi-level cross-check of data entered by humans;
- Communicate with clients in messengers via chatbots;
- Use predictive analytics to manage solutions based on statistics and data mining.
These are just some of many examples. While it all sounds good, how does it apply to real-world scenarios?
Automation of loan applications for legal entities
Before the introduction of the RPA system, one UK-based bank demanded that its employees collect and structure data from 9 different sources to check the creditworthiness of users applying to the bank loan.
Employees manually copied and pasted information into Excel files. More than 50 people were engaged in this routine work. And then six more people manually checked the entered data for errors. According to the bank's statistics, 20% of employees in the first group made 1-2 mistakes a day. At least one of the inspectors also missed one or two mistakes every day.
After RPA was implemented in the loans department, 30 out of 53 people were transferred to other tasks, and the number of errors in filling out applications was reduced by 10-15 times.
RPA: what’s behind the technology?
Now let's talk about what's hidden under the hood of popular robotic solutions. Typically, they are based on software that is easily adaptable to your particular needs, that connects via an API to internal systems and simulates all the necessary human actions. You can also add chatbots, text and human voice recognition and other elements to make your RPA as robust and intelligent as possible.
Each of the modules is a separate product, and all of them are controlled by a single platform that coordinates the work of RPA (it is a kind of conductor responsible for a coherent and clear performance of orchestra). The platform can launch robots on-demand or according to schedule, provide the operator with visual graphics on completed or failed operations, and also build detailed reports on RPA performance and impact on cost-cutting.
RPA solution can be based both on Open Source and proprietary software.
How to measure RPA value?
Statistics will help! According to the results of the HFS Research survey, many companies that have already implemented RPA technologies receive tangible benefits and gain a strong competitive advantage, namely:
- Overall increase in operational quality and efficiency;
- Increased speed of task execution and delivery.
In particular, the following results were obtained in the course of the survey:
- A decrease in human errors – 21%;
- Better quality of routine tasks – 21%
- Shorter time to execute a task – 19%;
- Reduction in tools and resources required to get the job done – 14%;
- Collection of historical data for process improvement – 7%;
- Streamlined processes as a result of the elimination of extra actions – 7%.
According to ABBYY, the introduction of RPA allows many organizations to reduce overhead costs significantly. RPA has the following benefits for business:
- 90% fewer manual actions;
- 5x faster speed of documents processing;
- 4x faster search for data from primary sources;
- 50% fewer expenditures on accounting documents processing;
- 3x faster to close a reporting period, and more.
What’s in it for businesses to implement RPA?
- Increased speed of all operations: RPA lets you do more without hiring additional staff or purchasing extra tools;
- Rapid ROI: the actual results are visible after the first pilot RPA project (typically after one month of RPA implementation);
- Easy scaling after the pilot implementation;
- Extremely high precision: due to the elimination/reduction of human errors in processes, many KPIs improve tremendously;
- Better security as a result of managed data privacy and improved predictability;
- Improved service quality;
- Better flexibility for any IT infrastructure;
- Strategic positioning of the company as an AI-savvy brand;
- Staff turnover reduction and better performance as a result of transferring to more complicated and intelligence-intensive tasks.
“On paper there had seemed no hitches. Alas! Forgotten were the ditches,” – you may object to the above benefits that look too good to be true. So let’s take a look at RPA risks and limitations.
RPA risks and limitations
To configure and reconfigure RPA performance each time your system or business process interface changes, you still need to build a human tech support team, i.e., add additional costs.
RPA requires a well-documented system and business process workflows. It is impossible to automate the chaos, so you will have to start with the initial audit of all business processes and their formalization.
Fortunately (or unfortunately), robots cannot replace a person completely (at least not yet). In any case, people are responsible for the automation work. Therefore, when implementing RPA, you should always have a plan B ready: how to manually support the work of the business in case of failure or forced modification of the robotic process.
It’s not the best idea to completely trust your robots, so you need to invest time and resources in constant monitoring of your RPA processes. It’s highly recommended that you appoint a specially trained person who will be the first to respond to possible problems in your RPA. Your in-house RPA analyst should be able to make informed data-based conclusions and help minimize potential financial or reputation losses.
Information security is everything. Robots should work in a closed software loop to avoid the impact of malware. For example, a fraudster can change the robot code to start transferring money to other details. Or your robot can be hacked and compromised to intentionally make mistakes in credit applications: while your sales will rise, your brand reputation will turn into bits and pieces. So trust the robots only to reliable professionals and organizations that will be able to ensure the security of your business processes during and after the implementation of RPA.
Any technology replacing the person becomes an object to Luddism, resistance to change. So brace yourself that some people in your or other departments will sabotage your RPA implementation efforts. To prevent this, you need to need to build an army of internal supporters and promoters of RPA, show the anticipated benefits and use those RPA advocates to educate other employees about the value that RPA will bring to their particular function or team/department. Show them how RPA can create new opportunities for their professional development or can help them move up their career ladder.
Tips for effective RPA implementation
First of all, even before the initial audit of your business processes, count the money, and try to answer the question of how much you’ll potentially save on robotization. Choose the right technology solution, calculate the ROI, and сreate beautiful graphs to get your senior management and stakeholders’ buy-in.
Then you need to understand which processes will be most effective to automate. Don't hesitate to involve RPA outsourcing providers and external consultants to help identify those processes: the view from the outside will be wider and more objective. Such companies and consultants have extensive experience in implementing RPA systems, and most of them have a set of ready-made solutions you can piggyback on instead of learning lessons and wasting money and resources on reinventing the wheel. In fact, you can build any RPA system from ready-made solutions, and it will be much cheaper than to design from scratch.
Once you've decided on the tools, don't throw everything at once. Start with one or two pilot projects, then analyze, make conclusions, and plan further. Don’t try to automate all of your processes at once! Start with the highest-priority ones and scale your RPA as you move ahead and measure its effectiveness.
