This is the fourth of a series of articles that I am writing in my blog. I intend to cover 21 technologies / ideas. Here is my publishing history.
What is Robotic Process Automation (RPA)?
RPA, also known as Software Robotics, is a software program. It is a Fat program, in that it runs on the end user's PC, laptop or a mobile device (thin programs are installed on Server and accessed through a URL). It is a sequence of commands which are executed by bots under some defined set of business rules and conditions.
Applicability of RPA
RPA is often the first step towards Digital Transformation. It helps automate repetitive tasks performed by humans in a business environment. The tasks that are best suited to RPA are those mechanical tasks that involve copy / paste and require the user to login into multiple systems.. RPA integrates with existing software applications without the need to develop complex integrations.
As shown in the diagram above, RPA is useful in case of the following four conditions.
- Business rules changing on a regular basis: In most of the growing companies, this is a reality. The management may decide to increase the credit limit to a group of customers to 50% for example, from the current 30%. While it may take a couple of days of manual activity in multiple systems to update this configuration, a one time tweak in RPA will ensure that this gets done in a matter of minutes.
- Changes in statutory and compliance rules: To catch up with advancing technology, Governments across the world are changing compliance rules more frequently than before. RPA can handle such changes more easily and accurately than manual operations in disparate systems.
- Overworked workforce doing repetitive tasks: A modern workforce performs a number of non-value adding (tech speak for useless, brain numbing etc) tasks that can be efficiently automated by RPA. Some examples are insurance claim form filling, new customer creation, new Vendor Creation, invoice approval workflow etc. Human involvement in such tasks increases the cost of training and retraining and also the cost of handling human errors.
- Disparate IT systems: A typical modern company will have an ERP System, a CRM System, an MDM System, a POS System, Multiple Approval Systems, Document Management System, Image Processing Systems etc. Updating each system for every change in business rule (see #1 above) is a repetitive task fraught with errors. In addition, such updates always tend to lag the business rule changes and will not move at the pace of business rule changes. RPA is the best solution to handle such situations.
Two types of RPA Bots
Based on the human involvement in activating them, RPA Bots can be classified as Attended or Unattended.
Attended Bots are like Virtual Assistants. They normally are installed on the user machine and are activated based on specific instructions from the User. These Bots are normally active at the User-machine interface in the front office.
Unattended Bots work on the back office tasks, doing the repeat tasks (Taking a backup, restarting the machines, running the overnight reports, scheduling tasks, sending early morning status reports) which do not require manual initiation. Normally installed on the company servers, they work on preset time following preset rules.
Uses of RPA
- It emulates human action. This relieves workers of tedious and repetitive tasks and helps them focus on value adding activities leaving the repetitive work to RPA.
- RPA is useful in doing high volume repetitive tasks across disparate systems
- Virtual data integration: Since RPA works across disparate applications, there is no need to develop complex integrations
- Information validation and auditing: Since RPA is highly accurate, the auditing becomes easier
- Accurate reporting:
- RPA reduces 'Technical Debt', which is the gap between rapidly evolving business processes and the slow to adapt IT systems
- RPA allows data migration from non-traditional sources like images.
Benefits of RPA
- Error free auditing
- Cost reduction: Reduction in cost of error rectification, and training costs
- Quick results: RPA can perform in minutes the tasks that take humans days
- 24/7/365 availability.
Business Use cases
Every industry has repetitive processes involving multiple disparate systems. Here are some examples.
- Hospitals: Patient Registration, Billing
- HR: New Employee Hiring, Employee Exit, Payroll Processing
- Insurance: Claims Processing
- Travel and Logistics: Ticket Booking, Passenger Registration
- All Industries: Customer, Vendor, Item Creation...
Myths
If you are like me, who, on hearing the word 'Robotic' in RPA immediately thought of those destructive robots seen in the movies Terminator or The Matrix and got terrified, do not worry. RPA is nothing of that sort. In fact it has nothing to do with Robots as we know them. It is a simple software program that helps people do their repetitive tasks more accurately.
You may know that RPA is not a Robot, but you still might think that RPA is very complex to implement. Perhaps my team may not be able to handle it, you presume. That is not correct. It is a wizard based solution that is easy to install, configure and maintain. Of course, some tech savviness is expected in these times, don't you think?
I can hear you telling, RPA runs without any human supervision, what if it runs amok? That is not correct. RPA needs regular human oversight to ensure accuracy in the face of changes to business rules, compliance requirements or technology. People are still in charge. Very much.
If you are thinking that you do not have the wherewithal to implement RPA because you are a small company, think again. Every company can identify repetitive non-value adding processes that can be automated using RPA.
Challenges
- Need for continuous oversight: Even a small change in the automated business processes will require re-configuring RPA. It is not a 'fill it, shut it, forget it' type of solution
- RPA can efficiently perform mechanical tasks. However it fails when the task involve cognitive inputs or decide among multiple options.
- RPA cannot identify erroneous data. So if the input data is erroneous, so will be the output delivered by RPA
- Depending on the tool used, there could be intense technical activity involved in deploying RPA. It calls for API based integration or simple connectivity to disparate systems
- Organizational change management challenges: Nothing spells 'Low Morale' in an Organization than the words, 'We are planning to deploy bots to replace you'. It is important to have a robust change management plan in action before initiating RPA. The plan should identify tasks that the employees could do as well as train them in their new roles. This is all the more important considering that automation is expected to replace 35% of manual jobs by 2035 as per some studies.
- Need for proper data management / governance policies: It is very important to have a well documented IT governance plan before deploying RPA. In one of the companies, someone changed the password without updating the bots and this lead to leakage of data.
- Scope creep: RPA will throw out a lot of data. It is very tempting to plug in a machine learning algorithm to make sense of the data and before you know, your RPA initiative has turned into a full-fledged AI initiative for which your company is not ready for.
RPA and AI
RPA is good at handling systems and processes that have a clear straight-forward and mechanical flow. On the other hand AI can improve decision making in complex, cognitive intense scenarios.
Here are some of the scenarios where they can work together.
In a hospital, RPA receives the email and moves it to a specific folder. AI can read the email, understand that it deals with stomach ulcer and move the case to the Gastroenterologist. At the doctor's office, RPA can pick the patient's data and schedule an appointment with the doctor and send a reply email to the patient.
As another example, RPA could collate data from disparate applications and push the collated data to AI to identify patterns out of the data
Or, RPA could pick up the machine performance data for AI to do predictive analytics based on trend analysis.
Potential is endless...
References:
- Automation – Robotic Process Automation - India | IBM
- What is Robotic Process Automation - RPA Software | UiPath
- Navigate RPA Journey & Steps - Automation Process | UiPath
- AI & RPA - New Level of Automation | UiPath
- Differences Between AI And RPA - When To Use Both | UiPath
- Scaling Automation Journey - Hyperautomation | UiPath
- How to explain Robotic Process Automation (RPA) in plain English | The Enterprisers Project
- RPA In 5 Minutes | What Is RPA - Robotic Process Automation? | RPA Explained | Simplilearn - YouTube
- How Does RPA Work? | What Is Robotic Process Automation (RPA)? | RPA In 10 Minutes | Edureka - YouTube
- What Is Robotic Process Automation (RPA)? | Introduction To RPA | RPA Training | Simplilearn - YouTube
- Robotic Process Automation: What is RPA? - YouTube
- What is RPA? A revolution in business process automation | CIO
- What is Robotic Process Automation (RPA)? | Automation Anywhere
- What is RPA? | How it Works, RPA Tools & Applications | Edureka
- Robotic process automation - Wikipedia
- RPA Tutorial: What is Robotic Process Automation? Application
- White Collar Robots: The Virtual Workforce | David Moss | TEDxUCL - YouTube
- Robotic process automation is killer app for cognitive computing | CIO
- Digital robots for everyone | Koichi Hasegawa | TEDxOtemachiED - YouTube
No comments:
Post a Comment