Proactive Response to Pandemic with Data Science and RPA
The COVID-19 crisis is changing the world we live in, requiring the need to respond rapidly and prioritize faster to support the global infrastructure across the industries. Sopra Steria, a leading European Information Technology consulting and digital services company of 45,000 employees, has forecasted the future needs- whether pandemic or environmental cases that will require predictive analysis to take preventative actions before a major impact occurs. The focus is to drive digital technology excellence in operations through intelligent robotic process automation.
The journey began in Jan. 2020, which is when the Sopra Steria RPA center of excellence initiated the project combining data science with RPA to create technology solutions that can prevent outages and impact to the end customers through foresight and insight. With the COVID-19 pandemic surfacing across the planet a few months later, the solution was quickly realized.
In the banking industry, it is necessary to streamline services for in-person banking and support of digital banking infrastructure–such as servers, storage, network, and virtualization resources, ensuring the system can scale and ramp up capacity and resources. The ramp up is based on the usage to optimize the load while providing flexibility to ramp down capacity or tasks as needed for cost management.
Before COVID-19, there was a balance between digital banking and direct/in-person banking. However, when there is a nationwide lock down where bank has reduced staff that results in long wait time, customers would prefer managing their financial and billing online. In the case of load scenarios where online becomes inaccessible or performance has decreased, the detection would come from the customer. Thereafter, a ticket is created, and an engineer goes through the entire process to increase infrastructure resources–such as server capacity which includes elements such as RAM, CPU, etc. Additionally, the process goes through a review, reproducible steps to prioritize, and finally, through the IT process to increase the resources. The complete infrastructure and virtualization are managed by IT supported in shifts 24x7. This entire process from creation to closure takes 1–3 days SLA depending on the criticality. By the time the infrastructure has the necessary resources, the customers have already experienced delay with exposure risks of site crashing due to limited bandwidth.
In the world of digital banking, the annual value of digital payments is anticipated to reach $726 billion while the volume of payment transactions is projected at $500 billion before the start of 2021. As a result, there is a high demand to ensure stability in the infrastructure every minute to minimize the loss of leads and/or opportunities while ensuring customer has a seamless experience.
The solution for digital banking transformation requires data science with AI and RPA for prediction and to take proactive measures before the issues reach the customers. Using AI and machine learning models, various news sites, newsletters, magazines, etc. are fetched and news are aggregated, trending from the prior and current days to predict upcoming workloads. Sentiments are gathered to map out the anticipated impact on the digital banking infrastructure. A machine learning model is leveraged to process unstructured data to structured data by breaking down articles and sentences that fall under a specified criterion to proceed with performing sentiment analysis on each sentence to get a score. Based on this, the unattended bot designed with UiPath Studio makes changes in the server console and increases or decreases the infrastructure as per the suggested AI element.
To ensure there is no impact to the customer, the bot executes the jobs during non-business hours and sometimes during service hours depending on the criticality state and duration of the forecasted impact. Throughout this automated flow, the load gets transferred from 1 server to another to balance the load accordingly. In the case where the criteria deem that the disaster has been lifted, the alert goes out to the configured bot, that normal operations has resumed to support in-person banking so that bot can take associated actions to decrease the server capacity which includes elements such as CPU, RAM, etc. based on the predictive criteria. This helps save unnecessary expenses that would otherwise be incurred had the server continued increased capacity when not required.
The lessons learned include:
• Harnessing the true potential of RPA by combining AI and utilizing various tool and technology to achieve great results.
• Solution designing that facilitates cross industry implementation and flexibility of solution customization.
• Ensuring the data model that identifies the news articles to be 99% accurate.
• If changes are made to the infrastructure for scale, it must not have negative impact that exposes greater reduction of resources.
• Load testing of the bot is required to ensure output generated is highly accurate.
The solution rolled out to Financial services during the pandemic has extended to other environmental emergencies such as floods, hurricanes, and earthquakes. This reusable solution can be leveraged across multiple sectors from government, public, manufacturing, to many more.
“This end-to-end automation solution developed in partnership with UiPath has helped our customers adopt a proactive approach and thus staying one step ahead. This solution is one of many, featuring an AI enabled technological advantage, that helps provide an array of benefits to our customers.” — Mohit Gupta, Director of RPA CoE, Sopra Steria
“We introduced UiPath 4 years ago into our business and to our customers as we are tool agnostic and wanted to select a world-class solution that meets our customers’ needs. Our customers have different requirements, infrastructures, &, budgets, in which we believe UiPath, the leading automation platform provides a simple, intuitive user-friendly experience, and cost efficiency to help with the digital transformation journey for our customers.” - Ankur Bansal, Head of RPA CoE, Sopra Steria