Adopting RPA to Automate IT Operational Processes
Automation is essential in organizations that performing digital transformation today. In a recent survey conducted by The Economist, more than 90% of global company leaders said that they do automate in starting the digital transformation process in their organization.
In Bank Mandiri, we start by automating rule-based, non-subjective, and repetitive tasks done by experienced workers. Considering some of those tasks can be done automatically, so they will have more time for value-added tasks to accelerate our transformation journey.
As a merged organization of four major banks, Bank Mandiri has many legacy systems and applications that are still operating. Most of these applications still use an old silos architecture that doesn’t offer easy integration through an Application Programming Interface (API). Whereas, the API is important for integrating various systems to get the process done automatically. With that limitation, we decided to use Robotic Process Automation (RPA) as a solution.
What is RPA?
According to IRPAAI, RPA is the application of technology that allows employees in a company to configure computer software or a “robot” to capture and interpret existing applications for processing a transaction, manipulating data, triggering responses, and communicating with other digital systems.
In a nutshell, RPA is a robot software technology that can be programmed to run a task or process automatically. RPA utilizes software programs called bots that represent a programmable or self-programming unit. The bot can mimic human-computer interactions, such as login website, filling forms, copy/move files, read/write/update data in a database, and other activities that commonly done by a human in completing a task through a Graphical User Interface (GUI).
RPA helps us integrate by leveraging the GUI on legacy applications that don’t offer an API. By going through the GUI, we don’t have to renew or make changes to several legacy applications because basically, each application has a GUI to interact with the end-user.
Why RPA?
Integration is the key to automating a process. In reality, to complete a process required interaction between multiple applications. Start from extracting data, manipulating, generating an output that could be a report, or also a prerequisite to trigger the next process. The activity involves several applications that could be different platforms (mainframe, web, desktop, or database). It takes integration between applications to automate the process end-to-end.
Based on our study, there are two possibilities that we can do:
- Renew the current legacy application with new technology and architecture.
- Adopt technologies that allow us to integrate across various systems without changes to existing applications.
In terms of business, the first option requires a significant investment, capital, time, as well as risks. Moreover, sometimes a process requires interaction from several applications that also need to develop. So, this is certainly not the best option.
Our choice is the number two option, which is adopting the RPA. Because RPA can access multiple platforms through the presentation layer, we can easily integrate one system to another.
The Infrastructure
We design the RPA infrastructure on the Mandiri Cloud VPC platform to ensure its reliability and availability. Since we use Unattended Robot, we need a server as the host of the robot. Therefore, we need a system that allows us to scale on the server when we need to increase its resources.
We have 8 Unattended Robots that projected to run 32 automation processes in parallel on the production environment.
Choosing Processes to Automate
We started the implementation around late 2019 by doing assessments on several daily operational processes in IT Operations, especially the end of day batch process.
In choosing which process will be automated first, we use the following matrix quadrant:
Based on the data we get on the assessment, then we compile the data to define the categorized of each process belongs to which quadrant. Afterwards, we will prioritize that process to implement it early, based on its categorize.
The figure below is the example of an end-of-day batch process. The workflow typically has interaction with some applications on multiple platforms. It doesn’t require cognitive effort because it fully rule-based processes. This kind of process usually takes 30–60 min on averages to finish.
We have been completely automated that kind of process for the pilot project. There are more than 20 batches of processes that must be completed daily, with a short window time of approximately 6 hours by 4–5 experienced employees.
The Benefits
a. Reduce Human Error
The processes run in IT Operations are generally a process that, although only performed once a day, have a high level of criticism. So, it has a high risk. By automating the process, we can reduce the risk of human errors caused by a lack of focus or external interference while executing the process. Unlike robots, it doesn’t get tired and has high accuracy.
b. It Optimizes the IT team to do more valuable work and increase their productivity.
c. Create the IT team opportunity to add new skills to automate the process.
d. Minimizes fraud risk, as it reduces user interaction to critical applications.
Future Implementation
We aim by mid-end 2020 we can automate all the end-of-day batch processes. Continuously we are still striving to optimize the utilization of robots by automating more potential processes both in the IT and also expanding to the business processes.
In the infrastructure side, we plan to implement a high-density robot deployment model where on one server can consist of 3–5 robots so that we can maximize the servers’ resources and reduce the time it takes to set up a robot server.
For the next stories, we will share our journey of implementing the RPA to optimize the business process. There are some cases in business processes that we still progress on implementation.
So, stay tuned!
