Banque Internationale à Luxembourg – Case Study
English version of the French publication from IT Nation in Luxembourg, November 2021
Going further with automation thanks to OCR
The BIL, Banque Internationale à Luxembourg, engaged Fujitsu Luxembourg and the Canadian company MagicLamp Software, to design and implement a document automation solution using IBM Datacap capture and optical character recognition (OCR) software and Robotic Process Automation (RPA) software from Blue Prism. This unprecedented project will not only further automation in the bank but also accelerate and improve the quality of several processes that were previously manual.
Still evolving just a few years ago, the adoption of RPA has considerably accelerated across many sectors. Banks, who have the willingness and the duty to digitally reinvent themselves, were quick to show interest in this technology.
RPA has been deployed at Banque Internationale à Luxembourg since 2016. It has been proven to be an effective technology for many use cases. However, as the BIL team moved the business toward “Hyperautomation”, they found there were some limits. One, in particular, is in handling unstructured data. Therefore, BIL explored other products that excelled at processing unstructured data and that could work together with Blue Prism’s RPA product. “At the end of 2019, we tested the potential of the IBM Datacap solution. IBM Datacap would enable us to extract data from documents in various formats from different sources and use RPA technology to compare them and input them into our applicative systems,” explained Franck Niatel, Automation Team Manager at BIL. “The test validated the interest in investing in this type of solution.”
A fruitful partnership
As BIL explored an implementation they looked for experienced partners in document capture and automation and who were also experts with IBM Datacap. This led to a collaboration between Banque Internationale à Luxembourg, MagicLamp Software and Fujitsu. Despite being located far away from Luxembourg, in Canada, the selection of MagicLamp was based on their strong capture and OCR experience as well as their numerous references. “Nevertheless, it was not easy to launch the project smoothly. Just as the project began, the Coronavirus pandemic started. For the bank, it was very difficult to measure the financial impact the crisis would have. However, BIL decided to continue the automation project which, after all, is intended to mitigate exactly this type of situation,” follows Franck Niatel. “Indeed, in this Covid crisis context, we observed that a lot of organizations, if they had not already done so, found themselves in circumstances that required them to accelerate their digital transformation,” agreed Steve Heggen, Head of Operations Automation at Fujitsu Luxembourg.
Once the decision was made, the teams worked entirely remotely. “Not only were we working from home, but we also had to learn about information access and collaboration between Canada, where MagicLamp is located and Luxembourg where the Fujitsu and BIL teams are located,” said Olivier Gourdange, Senior Sales Lead at Fujitsu Luxembourg.
Checking the coherence of data
To execute the project, the bank, MagicLamp and Fujitsu Luxembourg shared responsibilities. “Under supervision of the BIL Automation team, MagicLamp took charge of the IBM Datacap implementation and integration, making it applicable to all the documents the bank wanted to process,” said Benjamin Cormier, Datacap Developer at MagicLamp. “After deployment, we ensured knowledge transfer to make sure the BIL team who will use the solution masters it perfectly.” From Fujitsu’s side, the interoperability between IBM Datacap and Blue Prism allows for extracted data to be exploitable through automated processes.
Indeed, apart from its inherent capabilities, the selection of IBM Datacap was guided by the integration possibilities between the existing infrastructure of the bank and Blue Prism. “When it receives extracted data from IBM Datacap, Blue Prism can instantly detect if data is coherent or if it needs any complementary verification from an operator,” said Pedro Faria, Automation Team Leader at BIL. “This translates into gained time, while simultaneously improving and facilitating the necessary quality checks.” This combination of IBM Datacap to RPA was essential. “The objective is to give the operator more time to focus on complex decisions and to more quickly answer clients”.
Delivered in a six month period, the project has already shown impressive results. “In one month we have sent more than 3,000 documents through our first automated process with no major issues. We have undergone thorough testing to guarantee high quality,” observes Pedro Faria. Apart from these quantified results, which prove automation has been raised to another level, the team celebrated the adoption of the solution by the business teams. “You do not have a second opportunity to make a good first impression,” follows Pedro Faria. “We are happy with the positive feedback from the business users who report saving time.” The bank’s Compliance and Risk Management Services are also pleased: systematic and uniform data checking is faster and more trustworthy than before. They see significant improvement, important considering the numerous controls the banks must undergo.
Looking forward, the Banque Internationale à Luxembourg intends to continue to follow this path and go even further in their automation journey. The possibilities are numerous: anticipating issues with incoming documents by controlling their conformity, automatically sending the incoming documents to the appropriate departments, and using the natural language process (NLP) to identify the ideas and key information in a document are all candidates for future automation projects based on the foundation of IBM Datacap and Blue Prism. “We nevertheless need to be cautious when we launch these developments because, for us, the rule is simple: we want to deliver what we promised to our collaborators. That is how we will keep the trust they hold in us – and in these technologies”, conclude Pedro Faria and Franck Niatel.