FWR talks to Vestmark about the ways that AI is making an impact on the day-to-day tasks of wealth management and what the future could bring.

The day is fast approaching where an AI tool might not only transcribe a client’s conversation with a wealth manager or suggest the words, but it could also discuss an update or propose retrieving the facts during a meeting.

In all the debate about the use cases for AI (see examples here,  here and here), speculation continues about how deeply this technology will enter the activity that humans have performed for centuries.

Leo Schrantz, the new chief AI and innovation officer at Vestmark, a US-based provider of portfolio management and outsourced services, says that note-taking apps are gainng most attention now. But others may soon grab the limelight.

“To me the exciting thing there isn’t the time savings – it’s that you start to introduce AI into the advisor-client interaction. There are a lot of fascinating things that [it] can start to unlock. If you get clients comfortable with an AI listening and producing transcriptions/notes/action items, how big of a leap is it to welcoming the AI into the conversation as a participant?” Schrantz said. “Not necessarily to produce advice but to help retrieve facts, figures, go off and get an updated proposal [for the client] based on feedback in the meeting, retrieve details on specific holdings in a portfolio, etc.” 

The debate on what future uses exist for the US is “the trillion-dollar question for wealth tech,” he said. 

“I think we can look to the classic trends in wealth tech as a guide. Look today at the kinds of services that are the purview of the UHNW or family office segment where the services are only viable at a particular tier of wealth – AI will likely bring those costs down dramatically,” he said.

Schrantz spoke to this news service at a time when AI continues to be the hottest technology subject out there. AI capabilities include credit scoring and risk assessment; fraud detection and prevention; chatbots and virtual assistants; personalized banking and financial planning; algorithmic trading; customer relationship management; regulatory compliance; robo-advisors; and natural language processing (NLP). 

Schrantz has been at Vestmark for more than three years; he used to work at firms including Bain & Co. His specific role at Vestmark shows how significant AI is in the firm’s opinion. 

This news service asked what future AI use cases could come into view.

“One that I see active development on is how to manage private investments/alternatives alongside a more traditional equity/fixed income portfolio,” Schrantz said. “It’s a lot of paperwork management and wrangling disparate data and inputs to standardize in a way that you can get it into a traditional portfolio management software application like at Vestmark. You can throw bodies at that when you are dealing with $30 million-plus portfolio allocations – but technology is the only way to scale that in a $1 million portfolio.”

Nothing gets wasted
Schrantz said there is a risk of firms ploughing money into AI without thinking through what the use cases are and what they will get out of it. One risk is firms concluding, wrongly, after a failed idea, that there are no real uses cases at all, he said. 

FWR and Schrantz discussed how the term “robo-advisor” has largely fallen out of use today, whereas it was everywhere a decade ago. But that does not mean the lessons learned from the robo model should be lost, he said. 

A benefit of the robo-advisory period a decade ago has been a greater understanding of where value is created. For instance, robos have taken over some of the tasks of portfolio rebalancing, for example, and done so at scale, he said. 

“There is a possibility for AI to step in and do things where advisors have spent their precious time.”

A key issue for advisors is knowing what to do and how their judgement applies if it appears that AI has got something “wrong.”

AI can help show what the “workings” of an advisory process are, and this can build valuable data over time. 

“The Advisor Assistant we built is a conversation between the advisor user and the software. So that gives you an incredibly rich source of data on not just what a user is doing (`click stream’ data you might get, or log data, traditional observability data) but what they were trying to accomplish. And if they don’t get what they want you have an instant feedback mechanism to diagnose if something isn’t right. And it isn’t buried in some complex log file or code – it’s just plain language back and forth,” he said.