With artificial intelligence maintaining its position as the
On Tuesday, leaders from wealthtech firms that have become household names to advisors tackled the topic during a webinar organized by
The discussion was moderated by Ezra Group founder Craig Iskowitz and featured remarks from Andy Lientz, chief technology officer of Apex Fintech Solutions; Lee Davidson, chief analytics officer of Morningstar; Ted Denbow, vice president at RightCapital; Henry Zelikovsky, CEO of Softlab360; and Dani Fava, group head of product innovation at Envestnet.
Along with exploring how AI and machine learning are reshaping the industry, the session gave advisors tips on how to leverage these technologies to enhance investment decision making, optimize portfolio strategies and deliver personalized services to clients.
Davidson kicked things off by providing a quick update on "Mo," Morningstar's generative artificial intelligence chatbot that went live earlier this year across the Morningstar Investor, Research Portal, Direct and Advisor Workstation platforms.
READ MORE:
Dressed in a white-collared shirt underneath a branded company sweater, the
Mo is backed by the Morningstar Intelligence Engine, a platform that marries Morningstar's investment research library with the same Microsoft Azure OpenAI Service that powers ChatGPT.
When Mo receives a user's question, the engine identifies the most relevant content to feed OpenAI's large language model and construct a response. The response is then tested for relevance and responsiveness before being delivered to the user.
Morningstar officials said Mo was a popular guy out of the gate, fielding more than 10,000 inquiries in the first five weeks of his existence. Topics included questions on investing terms, saving for retirement, Morningstar ratings for specific securities, Morningstar's take on emerging asset classes and various market outlooks.
"Mo is essentially a digital research assistant using generative AI for financial advisors. So they're using it exactly in that way — as a front door to our research and analytics content," Davidson said.
But AI can't do it all. Davidson said one advantage human analysts will continue to have is access to the people who are managing these investment products.
"They can go out and do qualitative interviews and get kind of what I'll call exogenous information that's sitting outside of the datasets. That can be a really nice compliment to how a large language model could potentially utilize all the information we have within our datasets."
Lientz said one way Apex Fintech Solutions has been leveraging AI is in the realm of fraud protection and cybersecurity. Not only are they finding success using automation to prevent fraudulent transfers, they're keeping an eye on how would-be cybercriminals are using the tools to launch more sophisticated attacks.
To provide an example,
In mid-July, cybersecurity firm SlashNext published a
"One of the things large language models are doing is correcting grammar, and they're making [phishing attempts] look more realistic," Lientz said. "Where we really see it is people stealing accounts, not just stealing money anymore. We're putting in workflow tools to sort of help prevent that because we do know that the advisor needs to be part of the process.
"Customers are going to get trapped, advisors are going to get trapped. And so making sure that there's a second check on the human side before these interactions happen is going to be really important."
Denbow said RightCapital's approach to AI is still very much "wait and see" despite the red hot hype surrounding it. The financial planning software firm, which has
Ultimately, the talks center around one primary question: Should the software be making decisions?
"The clients are doing business with advisors because they trust that. There's a lot of emotion in money and management of that money. The clients want to know that the advisors are making the decisions, not the machines," Denbow said. "We're keeping a very close eye on it, as we think it's evolving. But we're not there yet. We have ideas where it may be of good use. But from that perspective, even though we're very new and very agile, we're a bit conservative from the AI side."
Fava, meanwhile, think's it is important to separate the discussion around AI into a pair of distinct categories.
"Because there's machine learning and predictive analytics, which I think a lot of our firms have probably been using for a long time. And now we've hit this tipping point where we're starting to figure out how to use generative AI to do things like create content, have conversations … and that's really why it's so popular right now," Fava said.
While not as buzzy as generative AI, Fava highlights the important role more established machine learning and predictive analytics applications still play in modern wealth management.
She pointed out the ability of those tools to do things like assess the risk of a client leaving a firm based on certain behaviors that they've engaged in, or predict which advisory clients likely have held away assets the advisor is unaware of.
On the generative side, Fava said Envestnet is experimenting in a lot of different areas, including finding ways to simplify overall operations.
"These systems are complex. All of us have complexities that have been built into the system through years and years of configuration. Can we unlock that institutional knowledge by loading all of that content into an LLM and making that available to all investment employees?" Fava said. "That is something that will add a tremendous amount of scale to client service and will lead to more consistent answers."
Denbow says he sees the potential of AI being used internally to help staff members tap into a firm's knowledge base and deliver consistent answers more quickly.
Still, he stands by the advice that when welcoming AI-powered solutions into your firm's ecosystem, consider them interns.
"Obviously, there's lots of busywork and lots of processes and operational efficiencies that could be pulled out of it. But it still needs a check. It still needs someone to kind of validate it and make sure that it's delivering what it's intended to do," he said. "So that's a bit of how we're looking at the ways we may use it.
"And if it could deliver coffee, we'd love to have it do that. But we think that a lot of it would be really just around compiling all that information and serving it up to our teams quickly and efficiently."
