on 01-03-24 12:26 PM - edited on 12-04-24 01:14 PM by Michael_S
Actives and Passives / Branch operations / Judicial Defense
In mid-2023, a law was enacted in Chile that obligates tardy debtors of child support to settle their debt with their own funds. This meant that official notices began to arrive from various family courts in Chile to all financial institutions in the country. The notice requires these institutions to report and seize the balances of the products, in case the debtors possess them. With the quantity of notices arriving day by day, the different departments were unable to keep up with resolving the cases, and they were accumulating in queue. The number of cases kept increasing, so it became necessary to find a way to automate the process, in this case, for one of the largest banks in Chile.
Discovery
As soon as the Law began to be processed, the Bank's team started preparing to implement some form of automation. However, I am unaware of these details. I know that this part of the project was carried out by individuals experienced in process mapping, but not in development using the tool.
Design
The bank's team knew they needed someone experienced in RPA development using SS&C Blue Prism. Being the Deloitte developer with the most experience in the tool, and having worked the most with the Bank's internal platforms, they decided to assign me the development.
The process design was unconventional from start to finish; I was only provided with a video containing the step-by-step instructions. There was no IPA, FRQ, PDD, SDD documentation (and if there was any document that could be similar, I was never aware of it). Due to the nature of the project, it was necessary to show results as soon as possible. Therefore, the design started from a Template flow, structuring the steps to comply with both BP's best practices and the bank's CoE, and developing on the fly.
Build
Broadly speaking, the process is easy to understand but complicated in its business rules, as it must be very strict when handling money.
A notice is received in PDF format, and the necessary data is extracted in the process; then the client is searched for in an internal platform of the bank, and it is checked if they have seizable products in different modules of the platform.
If they have a balance, it is seized; if they do not have a balance, the corresponding area is informed, and finally, a report is created.
Many business rules were created on the fly because there was no previous documentation. Rules such as, if the client has a higher balance in a cashier check or term deposit product, it corresponds to a referral to another area.
Quality Assurance
First, I conducted unit tests in a controlled environment, meaning all steps before making any kind of money movement were thoroughly tested to see how reliable the automation was.
Once all modules were tested, end-to-end tests were performed, making money charges and informing the business of the charges made. Finally, real tests were carried out with a resource on a controlled environment.
Delivery
Once we had the necessary confidence in executing as usual, the next step was to start reducing the backlog of operations.
To do this, we increased to 15 resources working simultaneously. Following Blue Prism's best practices and using environment locks when necessary, the process of deploying machines was quite fast. So the backlog of cases in the area was reduced in just one week.
With the success of this bot, we reduced to a normal daily flow of cases in the metropolitan region; now it was time to move the process to cover all regions of Chile. We repeated the step from Quality Assurance to Delivery, but this time with 25 resources working simultaneously, achieving exceptional results.
The constant challenge was the lack of documentation to guide the development. As mentioned earlier, I only had a video that didn't cover most of the business rules I had to solve on a daily basis. However, I wouldn't say this was the greatest challenge.
Initially, the described bot consisted of two parts that were ready by the delivery dates. However, during the presentation with the business team, we realized that a crucial step was missing, which involved adding a new platform to an already complex flow.
I informed them that automating this new step would take much longer than consolidating the two processes I had already developed. Despite initial resistance, the management agreed but insisted that everything had to be ready as soon as possible. It was frustrating to have two fully functional processes and discover at the last moment that the original plan couldn't be executed. Unifying the processes became an urgent requirement.
With much effort and sleepless nights, I managed to complete this additional process in just four days. Fortunately (or unfortunately), it coincided with a weekend, and by the following Monday, everything was ready.
At Deloitte, there is a culture of generating 'impacts that matter', and that slogan resonated constantly in my mind during this process. Bureaucracy in Chile is too slow, and there was a large number of people requesting the seizure of funds for the payment of child support, which generated enormous waiting times. Some of these delays were due to financial institutions not having the capacity to quickly resolve the received orders, which delayed people in their procedures. Personally, I can say that my mother requested this retention due to a debt held by my father, and I know how long it can take.
I have been developing bots for companies for over 5 years, and when a bot is in production, its results can start to be counted. However, this is the first time I have developed a bot that, while relieving pressure in the bank's areas, directly helps people who need child support money. For the first time, I felt I was making a significant impact.
In quantitative terms, in the metropolitan region, more than 5000 overdue orders were reduced in approximately one week, transitioning to a normal flow of between 30 and 70 daily. For the regions, more than 12000 were reduced in approximately three weeks, also transitioning to a normal flow of between 50 and 100 daily.
I do not have data on how much the bank may have saved but, during development, I was made aware of significant penalties for non-compliance with this new legislation, which were avoided thanks to this automation
For the community, I can say that in development, we will always encounter difficulties that are beyond our reach and that we cannot handle. However, what we can control, we must approach it in an organized manner and always follow best practices to minimize the negative impact they may have on our deliveries when events beyond our control occur.
I remember that when I started using Blue Prism 5 years ago, I relied heavily on the community to seek knowledge and offer help whenever I had time. However, over time, I stopped actively participating. My advice is not to stop sharing knowledge and to seek new ways to automate specific tasks. The Blue Prism community is a valuable resource that not all tools possess.
------------------------------
Daniel Sanhueza
RPA Professional Developer
Deloitte
America/Santiago
------------------------------
Hello Daniel,
Fantastic project, congratulations!
How long did the project take in its entirety, from discovery to go-live?
Regards
------------------------------
Leonardo Soares
RPA Developer
América/Brazil
------------------------------
Hello Leonardo! Thank you for your question!
The bank knew that the new law was coming, but I'm unaware of how many months it took them before they assigned it to me. I can say that it took me about 2 months to build the first functional prototype, and the second one took me about 1 month. The first one ran without many errors while I was building the second one, but when we realized that a significant part was missing, I recommended merging two flows into one. It took me 4 days. I was familiar with both flows, but it obviously took a lot of hours to set everything up.
Glad you've read this, regards!
What a use case, @Daniel_Sanhueza! Tight time constraints, risk of non-compliance, a surprise extra requirement and real human impact. I imagine this was stressful but incredibly rewarding. From everyone at SS&C Blue Prism - well done! I'm especially impressed by how this was able to be scaled up and down as needed.
Thank you so much for sharing, and congratulations on your Blueprints Debut badge 🙂
Hi Michael, thank you for your kind words.
Without a doubt, this is one of the most rewarding experiences I've had developing automations
Kind regards! 🙂