Financial Services Firms Find Multiple Use Cases for AI

Three executives, all finalists for WealthManagement.com’s CTO of the Year Award, recently discussed how their firms have implemented AI and what their roadmaps look like for the technology, during an executive forum at the 2024 Industry Awards in New York City.

“We’ve tried to inject AI into meeting conversations, meeting notes, which again, just saves a lot of time for our advisors—and actually, our clients are really interested in it,” said Paul Algreen, CTO of Cresset Capital Management. Many of the firm’s advisors were initially nervous and reluctant out of unfounded fears over client reaction.

“Clients love it,” he said of the results the firm is getting from its use of Jump AI, which is being used in a pilot program with Cresset’s advisors. The technology records and transcribes the meeting notes and put them into the firm’s Salesforce CRM where the inputs can trigger workflows and “get things moving,” he said.

“There’s just zero delays, so we can really delight the customer by turning around action really fast, and the advisors are loving it because they don’t have to type notes anymore,” said Algreen.

“We have been using AI for a long time,” said Inez Louzonis, a managing director and head of Platforms and Capabilities at Merrill Lynch, noting that the firm’s use stretches across several forms.

This includes chatbots, like Bank of America’s well-known Erica, which supports retail bank clients, and a similar chatbot on the Merrill side that serves financial advisors.

“It’s just anything the advisor or the client associate needs to do and just asking that question in simple, normal language and then getting that response back in a very usable way,” said Louzonis.

Tuppy Russo, head of client technology and operations at AllianceBerstein, said her firm rolled out its own chatbot to its sales organization eight months ago to answer questions about its products.

“This was an efficiency gain primarily in the alternatives space,” she said.

“We had a team that would answer these questions, ‘When is the next close?’

‘Is this fund closed for ERISA?’ ‘What are the liquidation terms?’ And the same questions were coming to the same team over and over again,” said Russo, noting how frustrating this could be, especially in light of her firm’s 24-hour service level agreement with its customers.

She added while this has been very successful, it required a great deal of testing prior to rollout to ensure accuracy, as well as monitoring and auditing on an ongoing basis to find questions that had not been answered, answered incorrectly or had been added.

“Anything we put in the hands of advisors, we want to make sure is 100% accurate because that, in turn, gets passed along to the client,” said Russo.

“Where we can be a little bit more experimental and where we’re testing with some other large language models, again, is around product and answering questions we’ve given to our CIO team and our investment strategists,” she said, adding that the strategists are accountable for the content on those products and there is less risk than an advisor providing incorrect information to clients.

Cresset’s Algreen described other ways his firm is employing AI, specifically machine learning, to analyze investment portfolios and generate customized summaries for its clients.

“You don’t have to have an associate or client service person trying to read all these reports and data in the pre-meeting time period; all the portfolio information is at their fingertips already,” he said.

Another area where Cresset has found a strong use case for AI is in the firm’s private investments and alternatives shop.

“We’re doing a lot of the same document ingestion [the firm has done for years], but the tooling has gotten so much better than in previous days of optical character recognition, where it was very error prone—it’s now being ingested straight into a large language model, it's just so much more efficient and easy for us to correct errors too,” said Algreen. The improvements have sped up data reconciliation for the Cresset operations team.

Merrill’s Louzonis described other use cases as well.

“We have about 95 events that are served up to the advisor to take action on, from tax-loss harvesting to data requirements or notifying them of accounts that don’t have beneficiaries on them.”

She said Merrill Lynch is also looking at note-taking and other communications analysis capabilities but has not implemented them yet. When it does, though, she said it will likely be far-reaching at the firm.

“We’re not nimble at Merrill; we have systems built upon systems upon systems and the same data in many places, so as but one example, taking that KYC document and putting that information everywhere it needs to be,” across the firm, she said.

Cybersecurity was another area where AI has a role for advisory firms, according to Cresset’s Algreen.

“The take-home message is that a year ago, I think the three of us would have said the attackers are winning because they have all the AI tools, and we don’t,” he said. “I think that’s changed a little bit in the last year and a half. I think some of the providers are really getting better at helping us use AI to identify threats, phishing, and vulnerability management.”

AllianceBernstein’s Russo said her firm was employing AI in cybersecurity, too.

“We’re lucky enough to have a centralized team across the firm with the CSO, data protection officers and a whole team behind that,” she said.

“Phishing, I think, is one of our top threats, and we’ve brought AI together with that too; we’ve employed AI in email detection, and we’ve been able to increase our email blocking by 20% by using those capabilities,” said Russo.

“I think if your cyber team is not using AI to combat threats you should be challenging them on that because there are now tools available,” said Algreen.

Comments

Plain text