“We were able to capture years of the bank’s domain knowledge essentially overnight.”
<small>Head of Client Analytics<small>
Since the 2008 Financial Crisis, Re:infer’s customer has had to adapt to reduced trading volumes, more regulation and shrinking margins. Cost-cutting exercises to date have delivered only minor improvements, putting great pressure on Post-Trade Operations to find further efficiencies.
Middle and back office processes in Post-Trade Operations were costly and ineffective, failing to support sustainable revenue generated in the front office. One of the clearest challenges was an over-reliance on email to get work done. Operations staff spent 70% of their day outside core IT systems, with 40% of that time spent communicating in Outlook alone.
Operations leaders knew deeper efficiencies lay in their emails, but lacked any insight into them. Traditional process mining tools couldn’t uncover the root-causes of operational breakdown, leaving staff in a cycle of inefficiency.
Operations staff spent 40% of their day in Outlook and 30% manipulating data
Re:infer’s Conversational Data Intelligence Platform provided a complete solution for gaining insight and driving value and automation from unstructured email data. The key differentiator was the platform’s ability to train the ML models using the bank’s own internal language.
The bank’s operations team started with three mailboxes - Settlements, Collateral, and Client Onboarding - due to their high traffic (50,000+ emails a year each) and considerable FTE resource. Re:infer ingested 12 months of email data from each mailbox. Its unsupervised deep learning models then analysed the emails, identifying root causes and presenting them visually to the bank’s subject matter experts in a zero-code user interface. This process was completed in just 8 hours of training.
Re:infer has analysed over 15 million emails across the bank’s Post-Trade Operations mailboxes.
Re:infer proved to be a highly scalable deep learning platform, quick to train while revealing the processes and relationships behind every team and client - in real-time. With these new insights, the bank identified bottlenecks, inefficiencies, risk events, complaints and exceptions, as well as client issues that led to further revenue opportunities.
Re:infer created an entirely new Management Information stream, providing a clear picture of which processes were highly repetitive and transactional. With this information, operations leaders knew what inefficient processes could be automated through integrations with Service Now, Salesforce and UIPath.
- Millions of pounds in change opportunities identified
- £617k worth of FTE savings generated
- 110 change opportunities identified from only three mailbox analyses