Talks Tech #44: RPA, AI, and Automation: How Women Can Opt for This Low Code Technology
Preeti Thukaram, Associate Director at Invenics Services Limited, shares her talk RPA, AI, and Automation: How Women Can Opt for This Low Code Technology. She discusses how Robotic Process Automation can create more efficient work environments and its implications for companies.
The story of automation began when people used to fill cans long ago. This was an assembly line infrastructure where people used to work like machines. This was replaced shortly by robots. Efficiency was improved. This led to the discussion about how we can also do business process automation. One primary reason that brought about this innovation and this thought process was there were a lot of clerks and people in the data entry sector who were not using their intelligence. They were working as robots. They had no idea what the data represented.
There used to be huge volumes of such data records spanning thousands of rows. They would just read it and type it. This means a low quality of work for the person doing this job. You continuously keep doing this kind of mundane job. You introduce human errors. Human beings have an attention span of about only 15 minutes or 20 minutes. You lose your attention span. You lose your efficiency. You lose your focus. The cost involved in repairing these issues when they occur was huge, and the inefficiency and inability to identify them were so huge. We thought automation could be a solution where machines can work like machines.
Humans can use their intelligence for validating, approving, and escalating. There might be some kind of infrastructure issue where a human being needs to step in. Otherwise, he could take a step back, and the robot, the programs themselves, can do a lot of work. This is how automation was envisioned. The main aim of this was to improve the quality of work, to improve efficiency and to reduce risks, and to reduce cost. RPA stands for Robotic Process Automation. These are software robots. There are several vendors today out in the market. Most of them offer a free academy, which is a learning center. There are courses offered at a beginner level, advanced level, and intermediate level where you can go through the courses. You can train yourselves, build these programs and see what they can do for you. If you're really interested, you can continue on to the advanced courses. Robotic Process Automation means building a software program that emulates the behavior of a human being in interacting with systems.
When a task is repetitive, which must be done daily, it is rule-based. What I mean by rule-based is I always log in to the same website, I always use my same credentials to log in. I always access the same information. I always download it to the exact location, probably a similar location with a date. Maybe the file might have a different date. It's rule-based, it's repetitive, it takes time, and it's inefficient. I can create a software program that can log in on my behalf, access the website, download the data, save it in my file, and then once it's done, I can start my daily job. It is going to do it very fast, and it's going to do it very accurately. It's not going to make any mistakes.
The difference between a normal program and an RPA is that a regular program works like software. Those software communicate between themselves as programs. They will pass some parameters. They will require some data. They will communicate through APIs. RPA interacts with the system like how a human being would do. It takes control of my keyboard, it takes control of my mouse or any of the interactive systems that I'm using. It'll go via the same graphical user interfaces GUI, that a human being would need to interact with the systems. It'll enter and capture the data in a similar way as a human being would do. This is called cognitive automation. We use the term computer vision. OMR is an Optical Image Reading. OCR is Optical Character Reading. These technologies mean software robots will use these technologies to read the characters and the images as a human being would read it. If there is text on the web page or a text on a particular system, it'll read it, recognize it, like how a human being would do, and then analyze the data the way you've written it in the program. Then it'll use it the way you interact.
Let’s say you're buying a mobile phone and would like to know how many mobile phones are available. What are the prices of these mobile phones? You would like to compare these various offers in the market. You would like to make a comparison of the features against the price. You would like to know if this phone is available in your region, if it is available in the color you want, in the specs you want. How quickly the delivery will happen. Is it a door-to-door delivery? Do you have to pick it up? There are so many websites and so much information. A person could type all this information from various websites into an Excel sheet. This person would've spent about eight hours doing this. An RPA can do it in 20 minutes or less if the program is written very efficiently and optimally. You create and train the robot by giving it various websites and you run the robot. The robot will go and access these websites on your computer and do something called data scraping. It cuts the data out, copies the data and then it'll download it. It can download it into an Excel file, a CSV file, a text file, in a database or wherever you want to place the data. In 10 to 15 minutes of your time, it is sorted, cleaned, and placed in a nice structured manner, which is human readable and then you have everything you want. Data scraping gives you a lot of options.
Using UI automation, in the case of data entry in the education sector, you have students joining every year. You want to enter their entire personal details into the data system. You need to assign coursework to these particular students. You can create the robot. It'll make the data entry by reading their applications. It'll make the data entry. Then it'll create their setup and their student login. It'll email them the entire end-to-end automation. Once the student submits his application, you can read the personal details. The robot will enter it into the system, create his log-in, assign his syllabus and send an email along with a copy to his teacher and head of the department. You'll be seeing it happen in real-time scenarios. This is done within a data protection environment. The cloud, wherever it runs, is protected. It's a secure environment. It can be authorized. The robot will have access to only the data present in that particular environment. You have full rights, rules, authorization, and authentication. You can set everything up so that the robot runs securely in a secure environment and accesses data as action policies. These days we have AI that can understand the nuances of communication, what is being said, how it is being said, to whom it is being said. Is it being addressed to the sender? Is it just a news item and a reference point? Is it something that action needs to be taken? This kind of intelligence can be built into the robot. Communication sectors and communication mining is critical because we don't need to spend hours reading through emails. Once you set up this robot, if an action has to be taken, it'll put that out and tell you parts of it can already be seen. They mine the data and go through considerable volumes to identify the action items inside it.
In process mining, an organization will run a huge business process. They'll have a supply chain. They'll have to procure materials. They'll make some product out of it. Then they'll have to market and sell this product. If they sell this product, they have to get revenue. They'll have purchase and sales orders associated with the sales part. They'll have to reconcile the two and they'll have to see if they make a profit. It may vary in bits and pieces, depending on the sector. If it's software sectors, they're slightly different. Overall, the process spans across departments. We can install a robot that does process mining. It understands the process that runs in this particular organization. It finds better pathways, rules, and more efficient rules to achieve the same end goal. We can use these robots to know where the loopholes exist, where the blockers are, and the ease of the channel, and create the shortest path mechanism. We call it the most efficient path.
Guest: Preeti Thukaram, Associate Director at Invenics Services Limited