on 03-10-23 09:59 AM - edited on 24-07-24 02:19 PM by Michael_S
Banks have different types of cancellation- Debit, Credit, Bulk, Corporate
During Covid, there was a large increase in card cancellation requests made from different regions. The requests were so high that the team were not able to meet their SLA.
The cancellation process involved downloading a set of documents; verifying the card details; cancelling the same & updating the database with set of details. As this was observed to be a repetitive task, it seemed to be ideal for automation, rather than hiring a team.
Discovery
Our initial step started with understanding the requirement. We asked the business to cancel a couple of cases manually to get an end-to-end understanding of the process. It wasn't enough to get a perception of the applications & the process, we also had to identify the differences in the procedure for different types of cancellations in different regions. All the interpretations were mapped into a Visio flow to plan our approach.
Design
The Visio map was ready but the business needed solution to be used for 5 different regions. Time was critical; the solution had to be in place as quick as possible. The Solution Design for this process helped me mark the differences in decision making of the bot across types & regions. As the process involved web applications, mainframe applications, Excel & Word documents, it was a complex process.
Build
Multi-bot architecture was the final plan of implementation with some basics to be taken care of:
We divided the process into 3 major parts: Loader Process, Main Process & Reporting Process
Loader Process
Picks all the cases & lined up in the work queue with required information for the main process to run.
Main Process
Interacts with the web, mainframe & MS Office applications; processing all the necessary business rules.
This process was designed in such a way that it could be run on 5 different bots at a time to process all 5 regions at once.
Reporting Process
This was a general process which did reporting for all the bots, with regards to automation rate.
QA
We started testing the process in chunks & also planned to deploy in chunks/phases. Initially we faced some challenges but we rectified the code & tried to get a consistent solution.
Deployment
As mentioned above the deployment was done in phases. The bot readiness responsibilities such as- Setting up the business applications on Production Bot, confirming the credentials, synchronizing the team mail box, etc. were taken care by me.
As an RPA developer, I think the common challenge for the automation team is to gain confidence & trust of business on the solution provided to them.
Also, other than that faced several challenges like:
The business had team of 30 working on this process, who were not meeting their SLAs due to the increased load. Now, they are a team of 8 who work to audit the reports shared by the bots. The business had planned on hiring 20 FTEs per region to accommodate the increase in requests so, ultimately, this resulted in a good FTE return. This process continues to be run today.
When you get an automation request, remember some key points-
------------------------------
Shreya Satish Udagave
Technical Architect | Associate Consultant
Northern Trust Operating Services, Pune
Email- shreya14khot@gmail.com
------------------------------