Automation Anywhere was introduced as an RPA tool to facilitate the digital transformation journey of the client organization. They wanted to eliminate a lot of manual tasks and provide quality work for their employees rather than them doing manual activities. They also wanted to make sure that a process was in place. They wanted a very low code or no code automation so that even business people could upscale themselves and automate. They wanted to free up their employees from such manual tasks and have a lot of bandwidth to do some other research.
In the use case related to the cancellation of a ticket, a lot of things have to be handled. When you cancel a ticket, the airline has to make an update at a lot of places. Only then you would get the refund. It involves a lot of processes. They could be related to the meal or something else. Even if you have not subscribed to any add-ons, all the checkboxes have to be ticked, and only after everything is cleared, you get your refund. Before automation, the refund process was very tedious. When they were doing it manually, any mistake could lead to incorrect transactions. For example, they might end up transferring a different amount. It was a very tedious process, and it was handled with the Sabre application, which was not very user-friendly. It was mostly a terminal-based mainframe application. It had a user interface, but a person with good experience was required for it. A fresher or somebody less experienced could not do it. You needed to know where exactly to go, and there were a lot of things that you had to learn to do this process. It was always done by an experienced engineer, and it was very manual involving navigating different screens. During COVID-19 times, there was a shortage of employees, and the airline was handling a lot of regulations, so this manual process became more problematic. At that time, we introduced Automation Anywhere. We were targeting only 50% automation, but we were able to automate 80% to 90%. It significantly improved the time to process a request and reduced the dependency on a senior engineer for the process. The robot could handle the process. We had programmed the robot. It was a very good experience. It helped the client navigate through tough times.
When you decide to automate a process, you try to understand the process in and out. You try to understand the exception scenarios and how to solve them. You prepare a process discovery document. While doing this documentation, you try to streamline the process because Automation Anywhere, as a tool, needs to be configured to automate that process. It forces us to ask a lot of questions, such as what if the file to be downloaded is not there. You come across all such things at the initial phase. That helps a person to restructure the process and fine-tune it more strategically. It saves a lot of time. It has plenty of features to do Excel operations. It has a lot of capability for querying. You can read data from Excel and process it using a query. You do not have to do Excel operations one by one. It is very quick. If a person is taking 30 minutes to do the operation, it takes only seconds or minutes to finish it. It provides a lot of efficiency.
You also have a lot of traceability. You have a lot of logs. You can see what happened after a process. When you have an audit, you have some traceability. A person doing a task does not always document each and every step. He or she would just do the task and finish it off. When you automate the process with Automation Anywhere, because the robot is doing the tasks, you get a lot of logs and a lot of traceability for auditing.
At the initial stages of automation, you also redefine and optimize a process. Instead of downloading 4 Excel files at different places, you try to standardize it. You give a standard name and a standard path. Everything is standardized, and a lot of focus is put on security and data privacy. We are very conscious of not saving it at any location that is accessible to anybody else. Subconsciously, we do all these checks. It improves the security and data privacy aspects. It also leads to process optimization. Like any RPA tool, it saves a lot of time, and it is efficient, and then logs are there as a good source of tracking.
Automation Anywhere is easy. It takes one to two weeks to introduce a non-technical person to this solution. They already have good domain knowledge, and they just need to know how to work with the tool. In about a week, we will be able to train them in basics. We can show them the interface. We can show them what goes where, but they will not be able to develop full automation or know about full RPA coding. They can get the basic knowledge in one to two weeks. Full training will require building a use case. It could be a small use case. If a developer would take four weeks, which includes development, testing, and other things, a non-technical person would require eight weeks. Only then a person will be able to learn completely and become a very good developer.
In terms of integration, there was SAP integration and Sabre integration. There was one more application. I worked only with these applications. If there is a component available, the integration goes smoothly. When you want to go for something customized and you try to build an integration, it is very challenging, but that is common with all the tools. When you try an out-of-the-box integration, it is very easy because they would have already tested that it is working. When you try to build something or customize it, that is when the challenges come. I have seen plenty of challenges with Automation Anywhere when you go for something that is not there in the package. You need a lot of support from the team and a lot of effort. The security restrictions from the organization were also stopping the integration. It was not happening. A lot of errors came, and we had to reach out to the network team and the security team. There were networking and security issues that had to be resolved.
Automation Anywhere has saved time and costs, but I do not have the metrics.