Problem
A large financial institution had rigorous duties to fulfil in ensuring transactions it oversaw were compliant with strict legal regulations. A complex system of robotic process automation (RPA) rules, combined with a team of human validators was achieving satisfactory performance in the task but at high cost and latency, while requiring large amounts of time and expertise to update and validate the process when regulations changed. The business wanted to explore more intelligent ways of automating the process such that it could be kept up to date with ease.
Solution
The client had a large number of labelled example records on which to conduct analysis and tests. Mind Foundry’s suite of data analysis and preparation tools were used to process signals of a variety of types, and produce a robust set of predictive instruments that could reliably predict the required class, while also accounting for regulatory requirements. The number of automated closes (not requiring a human input) was significantly increased, lightening the workload of the human team, and subsequently improving their accuracy due to reduced time pressure and volume of requests. Mind Foundry’s in-built testing, simulation and profiling framework allowed rapid testing of retrained models such that systems could be updated and deployed as soon as possible when new best practices and requirements were defined.
