How Merrill, Morgan Stanley and more use AI as a chill pill for jittery investors

There's plenty to worry about in the second half of 2024: The Fed has yet to budge on lowering rates, there's a highly contentious presidential election underway and a recent market drop sent fears of a recession surging.

During such volatile times, financial advisors tend to spend more time calming investors from making knee-jerk reactions. But with information changing so quickly, financial advisors have to be able to act faster than before. This is where AI tools like large language models have stepped in for firms including Morgan Stanley, Merrill and Morningstar. 

Merrill, for instance, uses AI to detect whether a client has opened its chief investment office's capital markets outlook e-issue, which advisors sometimes share with clients in response to market downturns. That was the case on Aug. 5, when the S&P 500 index closed down by 3% and the Dow Jones lost roughly 1,000 points. Merrill's advisors can see a list of which clients looked at the notice and reach out to them, ideally preventing investor panic. 

"On a day like today, it's really becoming useful because they know, 'These are the seven clients in my book or the 10 clients in my book that I really need to go contact today because they're probably going to be the most worried,'" said Nitesh Kadakia, head of innovation and advisor platforms at Merrill, during an interview Monday as the markets were falling. "So we don't want them to overreact. We do need to make sure we are connecting with and understand where participants might be." 

READ MORE: Market sell-off lets advisors tout wins from diversification

While the markets have calmed in the days following the sell-off, plenty of unknowns could trigger another disruption. Many investors are hanging on to the hope that the Federal Reserve will begin to cut rates, but its continued inaction could disappoint the markets, impair credit availability and potentially impact the economy. And while the presidential election doesn't currently have a direct impact on the markets, results could shift how investors manage their money when it comes to potential changes in tax law, for example.  

To be clear, tech leaders said AI cannot perfectly predict the future of such events. Morningstar, which launched an AI-powered chatbot called Mo last year, has been training it not to respond to any specific predictions like the next president, said Lee Davidson, chief analytics officer the investment research firm. 

READ MORE: Almost half of advisors worry about contested presidential election

"It doesn't give you much," Davidson said about testing out Mo earlier this year by posing questions about the next president. "It gave very generic, Associated Press style stuff. … And that's what we wanted it to do."

Morningstar, which thrives off of utilizing massive databases and analytics for the wealth management industry, needs its AI to be nonpartisan to ensure outputs are neutral but useful for portfolio managers and investors despite market and political upsets. 

"It's not required to have a political belief to use this information and make it useful, and make it relevant for investing decisions," he said. "We get tons of traffic whenever markets are volatile because people are uncertain. They need information. We give it. We try to do that in an unbiased way as much as possible."

READ MORE: RIAs need to be using AI tools 'yesterday,' tech leaders say

Another way AI is helping firms respond to market shocks is by providing response language quickly — language often backed by data analytics. For example, Morgan Stanley uses its large language model, AI @ Morgan Stanley Assistant, to help advisors craft the right content to respond to specific client needs during market disruptions and then deploy it fast.

"Advisors can then leverage subscription capabilities to deliver that relevant content automatically to their clients; covering hot topics like the election, market volatility or Fed news," said Andrea Zaretsky, chief marketing officer of Morgan Stanley Wealth Management and E-Trade. "Further, we can quickly adjust our content using machine learning based on client preferences to maximize relevance."

Zaretsky added that Morgan Stanley is also experimenting "with generative AI to enable our content marketing teams to quickly distill complex market and economic insights in digestible, approachable formats, like social media, emails or banners."