Morgan Stanley Hones Advisers with Machine Learning
Despite the results of several industry surveys, financial advisers should not be worried that machine-learning algorithms will eventually replace them.
At least not in the next ten years, according to Chris Kovel, head of analytics and data technology, wealth management technology at Morgan Stanley, said during a panel discussion at the Inspiring Innovation 2017 conference in Midtown Manhattan.
“When ATMs first came out, people thought there would not be any more tellers,” he said. “There are more tellers than before.”
The role of financial advisers will change as they gain access to more data on their clients than they ever had before.
“It took a good financial adviser to know when everyone had their 16th birthday,” said Kovel. “Now, they will know when their clients are about to have a life event and help them plan for those events.”
Morgan Stanley’s wealth management business has been using machine learning for the past to refine and personalize its research offerings.
“Back in the 00s, we used to carpet bomb our research to every single client without knowing what they wanted to read,” explained Kovel.
For the past five years, the asset manager has constructed client profiles regarding what research they read.
“By putting research consumers into groups, it also helps us recommend research that their peers have read, but they have not,” he said. “The same can be done with marketing efforts as well.”
The more data Morgan Stanley can gather on the behavior of its clients the more personalized an offering it can provide, according to Kovel.
In the past five years, consumers have grown comfortable with the targeted recommendations and personalization that machine learning brings, agreed fellow panelist Daniel Nadler, CEO of data analytics and machine learning provider Kensho Technologies.
“Jeff Bezos and Amazon figured out that they only needed a little bit of data in terms of personality and behavior to make accurate predictive suggestions around the things you want to buy,” he said.
Nadler noted one infamous incident in which an online retailer recommended to a 17-year old shopper that she might look at baby clothes based on her previous behavior. Her father was upset over the recommendation the retailer made to his baby girl and complained, wrote letters, and spoke to the media about what the retailer had done.
“It turns out that the e-commerce site knew more about his daughter than he did,” he said. “The retailer in question changed the algo so that once it knew the person’s age, it would only recommend certain things. It drives home the power of targeted recommendations.”
Such insights are why Morgan Stanley thinks its models are efficient ones, added Kovel.
He estimates that such technology could present thousands of investment ideas that financial advisers could share with their clients and bring a more personalized touch to clients with smaller accounts.
It is this new level of personalization that will keep financial advisors from being replaced completely by an instance of artificial intelligence, according to Kovel.
“As you get older and start talking about life, death, and beneficiaries, that will always happen using a financial advisers model,” he said. “We never see people having these types of discussions with machines.”
Seth Merrin has joined Neuravest Research, which uses artificial intelligence to construct portfolios.
AI/ML on the buy-side trading desk is a long-term program with short- and mid-term deliverables.
Marina Goche became chief executive at alternative data provider Sentifi nearly a year ago.
Early adopters have improved outgoing sales call volumes, call times and volumes.
Integrated access, directed machine learning and AI will become key in transaction cost analytics.