Recorded at the BNY Mellon | Pershing INSITE 2022. Media Center Powered by:
Advisorpedia interviewed Henry Zelikovsky, the CEO of Softlab360, at BNY Mellon | Pershing INSITE in June, 2022.
SPEAKERS: Henry Zelikovsky, Matt Ackermann
Matt Ackermann 00:06
Let's talk AI Artificial Intelligence. What does it mean for financial advisors, the industry and their clients? I'm joined today by Henry Zelikovsky, from Softlab360 To find out, Henry, welcome.
Henry Zelikovsky 00:19
Thank you for inviting me and pleased to be here. . .
Matt Ackermann 00:20
So Henry talks about what problem does AI help solve for, for the financial advisor and the industry.
Henry Zelikovsky 00:27
AI is a tool, it's a mechanism for extracting insights from a financial services data or any surrounding data data might come from non financial sources, but it could surround an advisor or advisor process or contracts advisor has the intention from our usage is to derive the meaning of particular content that is stored in databases and applications that provide data and quality of decision making support to the community.
Matt Ackermann 00:57
So, so really helps with decision making then, which is incredible. But give me an example. Maybe talk to me about how you're seeing people leverage this, ultimately to help improve their business.
Henry Zelikovsky 01:09
So one of our use cases is a company intelligent data, we work with them as a predictive analytics supplier of technology. We've for them extract the meaning of financial implication of cost savings and money management, let's say impact of illnesses, or traditional illnesses of COVID induced illnesses, where their expenses might be in the next series of life events by age group by demographic. And this prediction allows an advisor to decide how to manage your portfolio or particular constituents. So it's a tool to derive meaning and present the meaning in park and behavioral sense, as people might react differently to the fact that their life event might be impacted by illness. But hereditary diseases, and historical information we use to derive the data is a public data in part, and it's a private data, intelligent data bias. It's one of the examples, but it's a new entry into introducing how to manage a portfolio with the insight.
Matt Ackermann 02:11
And that's incredibly powerful, because every advisor would love to know how long do I have to make this nest egg last for a client, right? That's really powerful.
Henry Zelikovsky 02:21
It is. And this is a dimension that intergender introduced. And they came to us with the possibility of finding if that would be possible. So this is an example where AI serves as a mechanism of looking at the vast vast number of data sources. And through the mechanism of prediction. There are various technologies and various methods. We particularly use Bayesian, which is behavioral science concept. We've been using that since 2014, and learned number of mistakes and number of improvements and how derive the meaning of the data. And it's the meaning that we're after. And once you present the meaning of the data to the person who can apply it, this becomes a valuable information driven tool.
Matt Ackermann 03:02
Very powerful tool, but can feel a little overwhelming. I'm sure it's sometimes for someone who's new to this, if I'm an advisor, I'm overwhelmed by this idea, but I'm intrigued. What's a way to get started so that they can start to see the value without being overwhelmed.
Henry Zelikovsky 03:16
We provide approximately eight to 10 weeks of what we call machine learning training exercise. In the traditional sense of machine learning the technology has to learn or train itself. On what data we provide it. We offer data science and mechanisms to filter what we call statistical noise. Something that is not palatable to learning will learn from the data from a number of techniques clustering data or segmenting the data in number of ways. We yield the result we discuss the result with the party who is asking us business questions. So we can derive the common agreement that this has a historical learning curve value. Once we get into accuracy above 94%. We feel they've learned enough and we can take new data and continue evolving this prediction. So from the advisor perspective, or any user of our technology, it's more than conclusion of proof of concept or conclusion of a learning period where they can see the result meaningful to them. They will not need to do anything about how we get there, but they can see what the next step would be getting that result that's meaningful, then, very powerful.
Matt Ackermann 04:18
Henry, thanks so much for making time today. Amazing insights. Thank you. Thanks for Advisorpedia. I'm Matt Ackermann.