DTCC’s Gordon Sands, in FOW (Futures & Options World), describes three common misperceptions about the use of robotics in the financial services industry and recommends ways financial services firms can best integrate this technology into their operations to generate efficiencies and expand organizational capacity for higher-end activities. Access Gordon’s article, along with other related content here.
DTCC is among a growing number of firms in our industry that have begun integrating advanced technology into selected processes and products. These applications allow us to test and refine new approaches to automation and increase our capacity to deliver higher-value data to internal and external clients, and the industry at large. As we gain experience with robotics, AI and machine learning, we expect to expand their use across business lines and functions.
Here I share some of our experiences putting robotics to work in the DTCC enterprise.
Finance Department’s Invoice Reconciliation Project
One of our earliest pilots of robotic process automation (RPA), this initiative has transformed a sampling-based reconciliation task into a comprehensive audit of all DTCC outgoing monthly invoices.
Our Finance Revenue Cycle team uses two different legacy systems for billing, which requires invoices from both systems to be reconciled at the end of every month. Because the group lacked the human capacity to reconcile each invoice, it had been sampling 10% of the total as a broad check on reconciliation rates. The sampling procedure also required the Revenue Cycle group to prepare complex monthly reports for Internal Audit.
By converting invoice reconciliation into an RPA application, capacity constraints disappeared such that every single invoice was reconciled and the manually produced monthly audits supplanted by the robot-generated reconciliation file.
Using specialized RPA software, it took only three months to go live with the new process. Staff bandwidth has been amplified, freeing up people to focus on analysis and troubleshooting of those cases that fail to reconcile. Clients have also benefited, through more accurate invoicing.
Mutual Fund Services Profile II Database Enhancement
DTCC in June 2017 implemented an advanced automation initiative incorporating AI and machine learning to strategically enhance Mutual Fund Profile II, a central repository for the mutual fund industry that maintains prospectus and operational processing rules for some 27,000 mutual fund securities, giving it greatly expanded data capture and self-learning capabilities.
This project, implemented in collaboration with Kingland, a firm providing AI-based enterprise data management services, leveraged recent advances in cognitive technology and data mining. By applying an array of AI-driven enhancements, we automated the data sourcing, boosting from 4 million to 5 million the number of data points covered by the database, and streamlined clients’ collection and sharing of this data.
The enhanced data-sourcing engine now captures higher-quality information and with greater frequency, and pushes it out to the industry faster. Its AI capability allows the database to continually improve by responding to users in real time. When discrepancies arise as a fund updates its data points — such as minimum/maximum sales charges, underwriting fees and social codes — the application automatically generates a notification prompting a data review, preventing the system from being updated with incomplete information.
By removing the typical burdens of data management, this project created capacity for data analysis, a big win for our organization and our clients.
GTR Onboarding Project
Our GTR onboarding project, an RPA pilot that is still underway, automates some of GTR’s complex onboarding workflow. Because this process enhancement touches other DTCC applications, it requires careful design. GTR, our Global Trade Repository business for over-the-counter derivatives, operates in multiple regulatory jurisdictions and has a 6,000-plus client base spread around the world. This global reach along with the highly regulated nature of derivatives reporting helps make GTR’s onboarding process detailed, time-consuming and highly risky if mistakes are made.
In the pilot, a new client submits onboarding forms, which are reviewed by a GTR team member and moved into a Salesforce queue if the information is deemed valid. At this point a bot processes the application and either funnels it into an exception handling queue when some data is non-standard or questionable, or designates it as successfully completed.
Because the onboarding bot touches other systems like Salesforce, even small changes in those other systems, such as a popup warning of an error condition, if not coded for, can cause the robotic process to fail. This project has helped us improve our design of RPA programming to consider all the application behaviors the process will touch. DTCC has discovered that developing standards and promoting them through an open-source model provides for both creativity and structure as the bots are developed.
Read more examples on the transformation of the industry’s IT profile here.