Robotic Process Automation

Robotic Process Automation

Have you come across terms called RPA, Workflow automation and AI. In this era where businesses are transforming from Business process outsourcing to Digital Process Management which presents new opportunities to reduce cost, staffing and delivering improved performance.

Robotic process automation (RPA) has played a key role in this transformation. So what is this RPA? RPA is a set of customized instructions that execute repetitive tasks. It emulates the work flow of human based repetitive tasks (certain operations) but is not to be confused with Artificial intelligence. RPA when pipelined into a series of various process executions is known as work flow automation. Most of the processes in the modern industries are supported through RPA whether it is automated email generation to attend queries, extraction of files from databases, software testing, Web scraping and so on.

Evolution of RPA

While the roots of the term ‘Workflow automation’ date back to 1920s during the industrial age, this term became more familiar in 1990s as RPA emerged supporting the mundane manual jobs such as information collection , data entry, transformation to the database and notifying the corresponding employee.

RPA is not something very new yet has not reached its full potential yet. A foundational question arises when thinking about what should be automated and what should be done by humans. Usually while thinking about automating a process the company usually thinks what kind of repetitive task has to be mitigated from human oriented execution to complete automation which involves the least intervention of humans, at-least at the operation level. This question has been in the industries for quite a long time.

Traditional automation which is achieved through writing APIs for each operations that connect front end, back end and storage can take several months to be implemented as it is developed at the infrastructure level. On the other hand RPA can automate high volume of repetitive tasks in the front end with a Rule based approach. RPA is tool based and RPA tools allow users to design and deploy software robots that can mimic human actions. These tools also utilize pre-defined activities and business rules to autonomously execute a combination of tasks, transactions, and processes across software systems. RPA can deliver the desired result without human interaction. On top of that RPA is quite faster to implement as not much of real coding is required and automation can be achieved with just drag and drop options with the help of RPA tools. This also makes RPA cost effective in terms of computation and manpower and through a little training such that even a non-technical person could pick up RPA.
Evolution of Automation

Seeing where the RPA fits in

" RPA lets bots optimize your business process, leaving the humans more bandwidth for the high-value needs."
Fitting RPA depending on the process 

RPA is not a replacement for the human customer-service work-force. It is ideal for tasks that require no human intervention, also often referred as unattended tasks. Plenty of tasks require a human connection, but in many cases at least a portion  of work can also be automated  such kinds of tasks are called-Attended task

One direct implementation of attended tasks are Desktop bots which automate the repetitive tasks of the desktop. It can also provide employee sitting at the in front of screens with information and guidance when and where it is needed. On the other hand there are Server based Robotic bots that do not require any human intervention and can work in an unattended way.

RPA Management

A digital workforce no matter how intelligent needs to be managed and monitored. You would not send a group of human employees onto the job without managing and monitoring. It is not a matter of lack of trust but rather ensuring that activities and processes are carefully orchestrated, coordinated , governed and executed. So certainly a part of management is regular evaluation of job each worker is doing, so any issue is addressed promptly and effectively. Similarly a good RPA solution requires some management that ensures that the robots are performing their operations flawlessly and punctually.

This includes a centralized management control room overseeing the work of automation bots. Some of the tasks handled by the control room is task queuing, collection of data for the process, monitoring on overall systems health and livelihood and allocation of robotic resources that matches the demand at that time.

With the control room the RPA system is quite reliable and scalable, and there is complete control and visibility with a dashboard that provides insight on operations. Processes can be prioritized , scheduled , start, paused and stopped through this.

Bringing RPA to your Organization

So you have a company with many operational processes which could possibly be handled by RPA but you wonder where to start and have some reservations and doubts whether will RPA bring a positive Impact? Do I need RPA ? Will it be difficult to migrate from Traditional approach to the Automated approach? Well it can be tough making such decisions initially but to start ,  experimenting with high volume but less complex tasks is advisable. Something that can be developed quickly and deployed successfully will be received positively both internally from the employees and by the client.

Few lessons you need to learn from RPA adapters step by step is as follow:

1. Prepare the organization :

  • Positive Alignment of all stakeholders from top management, process owners and technological department.

2. Ensure process Viability :

  • RPA is most effective with rules-based and consistent processes.
  • Developing an internal process audit to gauge the viability of RPA solutions before beginning RPA implementation.
  • RPA is not end of all human interaction , organization must account for exceptions and manual decision making.

3. Understand the Impact of RPA :

  1. Pilot phase processes should be well suited for RPA and should showcase a high-impact , tangible business benefit in terms of time and cost to promote RPA in broader organization.

4. Develop Capability :

  • Organization must develop internal capability  to manage and maintain the RPA environment.


As a decision maker or a manager in your company you would definitely ask yourself this wise question and the usual answer any vendor or solution analyst will tell you is "it depends". There are several factors your process automation depends on but with clear understanding helps to estimate the required time and need for such adventures.  The initial idea to automate a process will evolve with time and new requirements and improvement will come to your mind but rather than being scared of this ,it should actually encourage the company to invest more time and energy. To begin with sketch out as many factors as possible to describe your process automation for it will allow you to bring your idea of eliminating repetitive tasks to life.  This starts with measuring the complexity of the process.

To know how long will it take to automate a business process, its complexity aspect is essential. The complexity is usually measured by asking "How many applications are involved?" and here application are a set of tasks that the robot switches while executing steps (automation) throughout the business process . As the bot is mimicking user behavior it will focus on a single application at any given time.

Here is a step by step study to help you visualize an application of RPA.

  1. Process is initiated on reception of email.
  2. Read the email and download its attachments to a shared folder.
  3. Sign in to the CRM system from where data has to be read.
  4. Open a web-application to input and read data
  5. Open an ERP system and enter data which was read
  6. Generate a summary and store it in drive
  7. Send an email back to sender.

The diagram shows a simplified example of your automation steps and the respective dependencies between them

Once it is clearly what applications are to be automated, it becomes effortless to classify your business process and implement it .

How to Measure RPA's Success?

To measure success you have to look at the benefits you are looking for, which in RPA one can quantitatively  matches that were used in measuring the savings and investments. For instance, if you are looking for productivity gains, you should measure those areas to ensure RPA implementation and processes are yielding the expected benefits.

Man hours are the unit of measure in project management to calculate efforts needed to complete a task. It helps to identify the cost of project and manpower handling based on expertise and contribution.

*1 man hour = work completed in an hour of uninterrupted effort by an average employee*

The total man hours per task is obtained by multiplying the number of people allocated to a task by the total time it takes to complete it.

Example if 10 people are working at a task and take 90 days to complete it

Estimating 5 effective working hours per day, the total effective man hours is obtained from multiplying the following:

10 x 90 x 5 = 4500 hours

It is human nature to lose focus and deviate therefore even if an employee works for 8 hours his effectiveness is assumed only to be 5 hours for this example.

As a decision maker you definitely want to know how many man hours are required for a task and the cost associated with it. This helps you compare how effective RPA can be in saving your man hours and cost.

Case Studies Of Successful RPA Implementation Across Industries.

Healthcare technology firm saved 423,500 man hours with RPA.

A global health-tech company with its presence in nearly 95 countries automated around 300 processes aiming to save 900,000 man hours in operations of finance, document management, and analytics. Over the course of year working with RPA vendor the company was able to automate 140  processes with 200 bots working 3 shifts a day. As a result 423,500 man hours were saved cumulatively.

A bank managed to bring error rates down to (almost) zero

A bank in India deployed about 750 RPA bots that handle close to 2 million transactions per day.  As of now no cost or time savings have been reported yet, it is astonishing to see that human error has gone down significantly.

Replacing human agents

A major courier-service company provides an example of how RPA eliminates human agents completely out of loop in certain situations. Previously , call center agents handled claims and scheduled package redelivery. Claims would take weeks to resolve and and redelivery scheduling was inefficient. Now, a robotics solution reviews claims , checks refund eligibility based on business rules, transfers payments and notifies customer by email without any agent intervention. And automated redelivery scheduling lowers the volume of calls and errors.

Utility sector

In utility sector, an oil and gas multinational handles around 15000 change of address request every month. The existing processes were error ridden and time consuming. The bots came to rescue here. The human employees are still involved but robotic solution helped in merging all into single interface for gathering customer data. The robots handle creating new accounts in the CRM and the average time for change request dropped from 11 minutes to one minute.

We see how success of RPA is measured in terms of time saving, cost saving, error reduction, and effort reduction.

Leading Tools For Implementing RPA

Uipath is user friendly and easy to use and the cool thing about Uipath is that it has a community edition which allows everyone to begin with learning RPA, while the enterprise version is paid and has increased features. It provides a visual designer which speeds the implementation of a project as the whole flow of control is visualized. UiPath can handle complex processes and can be used for businesses of any size . It can be hosted on cloud and implicit terminals.

BluePrism is highly reliable and scalable tool with advanced features. It is trusted across the globe. One needs to have some programming skills to work on Blue Prism, which works on any platform and with any application One drawback is that it does not provide a free community version and to learn BluePrism one has to pay.

Automation Anywhere is an excellent tool in the market to develop RPA bots. It has both cloud and on-premise services and is best suited for medium and large businesses. It offers a 30 day free trial version after which one has to pay to use it. It has high security through authentication, encryption and credentials. It can quickly automate complex tasks and provide script-less automation.

From RPA to Hyperautomation -  A Data Driven World

With growth in RPA the growth in hardware industry and silicon architectures (Moore’s law) produced powerful hardware that can perform billions of computations per second leading to implement of more mathematically complex but accurate algorithms that can replicate certain human tasks involving decision making. The only problem was that these algorithms need to be fed with data else their existence would make no meaning. Co-incidentally with internet being used widely a very large amount of data generation occurred everyday giving companies an opportunity to utilize this data to test it with algorithms. This gave rise to more research and implementation in the field of mathematical modeling i.e. “machine learning and deep learning”. This gave rise to more disciplined study known as data science. This whole process focuses on how decision making without direct human intervention could be achieved through data dependency. Such decisions are data driven and When RPA is integrated with data science it results in called Hyperautomation.

In hyperautomation the bot not only follows the rule based instruction but can also comprehend the meaning behind each activity and make intelligent decision. Example in a hyperautomation bot can read through file, grab the context with natural language processing and deliver a short summary on it before making a  wise decision. But this technology is quite costly in terms of resources to be developed and require tons of data and hardware capabilities to be developed. But the advantages are immeasurable.

Some implementation of Hyperautomation are as follows

  1. Optical character recognition : Extracting contents from non standardized files
  2. Intelligent chatbots - Attending customers, precise query handling.
  3. Claim handling - Processing claims in insurance without any human intervention.

Continuous Improvement

RPA is by definition, a process improvement. And those who are aware about the business of process improvement know that improvement is never a done deal, there is always room for more improvement. The best RPA tools are not just for building automation but for identifying and planning before the build as well as monitoring afterwards.