“I do see blockchain coming into play and having an impact on the industry, but I believe that it will probably be a year or two before we see a big shift in that direction.”
Machine learning used for risk assessment
Discussing the ways to make wealth management more efficient, Larry mentions that machine learning may be applied to all areas, whether it is portfolio optimization, chatbots, or risk assessment.
“On the risk side we’re already doing that. We’re pulling data from other FinTech providers to pre-populate the questions and update the scores. We also have 100 variables built into our models based on life events that when triggered will automatically update the risk capacity score. The more data we pull, and if it fits that variable, the more we’re learning about that client.”
Larry claims that unlike a credit score, which is based on five common factors, Totum Risk uses over 100 variables for risk assessment.
“Are they married? Are they single? Are they divorced? Are they separated? All four of those are different, and they’re going to have a different risk assessment that will update their risk capacity score.”
Larry considers that gathering more data, having more users in the system, and having more integrations, together with machine learning, will all be great for financial advisors and individual investors.
“We tested this with millennials and they loved the idea of having the risk assessed for them when they sit down with an advisor, instead of going through the old-school, typical psychological risk preference questionnaires, which really are meaningless.”
According to Larry, Totum Risk has created a system that enables any investor to just type in their name and phone number to see their risk score.
“But not just the risk score. We’ll have a full narrative explaining how we got to that conclusion. And that’s all be based on machine learning.”
However, Larry expects a significant challenge to arise in machine learning implementation. Academic quantitative models need to be built for multiple scenarios, where the data is updated and run through the models.
“If you don’t have deep academic quantitative models with a tested algorithm then you just have a bunch of data.”
This is why it is important to make sure that information in the data sources used is up to date. Larry says that think if data is not updated at least annually then it’s old data, and the result is not going to be accurate.
Future expectations
Larry believes that in the near future we will see further growth of AI and blockchain in the financial advisory and wealth-management space.
“But again, the products that are using AI have to have those quantitative models and algorithms set up that are true and accurate.”
The biggest challenge Larry sees is in finding people who have experience working in the financial industry as well as a deep technology background. As an example, Larry suggests that AI models should be built by PhDs with relevant expertise in the industry.
Interviewed by Vasyl Soloshchuk, CEO and co-owner at INSART, FinTech & Java engineering company. Vasyl is also author of the WealthTech Club blog, which conducts research into Fortune and Startup Robo-advisor and Wealth Management companies in terms of the technology ecosystem.