If you're a senior banker spending hours each day working Zoom calls with clients, you might not want to be told how to handle client interactions by a machine learning algorithm.
A distinguished engineer at Morgan Stanley suggests this is a little closed-minded.
Speaking at this week's NVIDIA conference, Richard Huddleston, a distinguished engineer at Morgan Stanley in New York City, said deep learning algorithms and anomaly detection can be used for - among other things - improving client interactions.
It's a question of collecting data on how client interactions evolve and their outcomes, said Huddleston. "You would be noting down the way that the client contact is initiated, you would collect detail on the nature of that client contact, what actually was discussed and then the outcome of the client contact in terms of whether it was a positive or negative outcome." When sufficient data had been amalgamated, Huddleston said you would apply deep learning anomaly detection methods to look for anomalous behaviors.
When those behaviors lead to positive outcomes, Huddleston said you would seek to replicate them. When they lead to negative behaviors, you would "clean up and remove those types of interactions."
Huddleston's team specialize in anomaly detection and work on applications across the bank. Morgan Stanley has a "robust" and "multidisciplinary AI team," he said, adding that anomaly detection can be used for everything from client security, to improving corporate culture, fraud detection, anti-money laundering risk, risk mitigation, regulatory compliance, quant finance, cyber-security, outage prediction and detection, and IT space and power optimization.
Have a confidential story, tip, or comment you’d like to share? Contact: email@example.com
Bear with us if you leave a comment at the bottom of this article: all our comments are moderated by human beings. Sometimes these humans might be asleep, or away from their desks, so it may take a while for your comment to appear.
Photo by Hello I'm Nik on Unsplash