Qraft US CEO: Using AI to Improve Investment Decision-making

Artificial intelligence is being used to automate and transform back-end processes of financial services firms. But when it comes to wealth management, the adoption of AI has been slower. That may be changing.

Robert Nestor, U.S. CEO of SoftBank-backed Qraft, discussed with AI Business how AI is changing wealth management. A former executive from Vanguard and BlackRock, he explains how AI can be leveraged to supercharge human decision-making -- helping advisors especially during volatile markets. (Nestor has since left Qraft and is now an adjunct professor at Drexel University in Philadelphia, Pennsylvania.)

What follows is an edited transcript of the conversation.

AI Business: What does Qraft Technologies do?

Rob Nestor: Qraft Technologies is the classic fintech, investech firm that's really at the intersection of bringing artificial intelligence processes into investment and decision-making, whether that's asset allocation, security selection or trading insights. We want to be the driver of the enablement of artificial intelligence (in wealth management) across the board over the coming generation.

AI Business: What pain point or problem is Qraft solving?

Nestor: I would say maybe less of a pain point and more of an opportunity. As we all know, artificial intelligence has changed our world dramatically. … But the asset management business, and specifically, the process of investment decision-making, either at the bottoms up level with securities, or the top down level really haven’t participated in the use of artificial intelligence.

We see the real opportunity of taking that very powerful human intuition that is a big part of investment decision-making and leveraging that with machine learning, deep learning other artificial techniques to dramatically increase the scope, scale and speed of analysis.

We think the adoption in the investment decision-making portion of asset management is inevitable. We are pioneers there and we want to be a key enablement of that over the next 10 years and beyond.

AI Business: Where is AI being leveraged today in finance?

Nestor: For probably 10 years or more, AI has been an integral part of a lot of trading functions in our business. In fact, the origins of Qraft were actually on the trading side and that is a meaningful business for us in Asia today.

More recently, artificial intelligence played a big role in operational efficiencies in the asset management business, not just traditional processing and bringing scale to that, but also helping decision makers like wealth advisors understand and analyze their book of business at a much greater scope. A lot of the more recent developments in this space have been around that.

But it really hasn't penetrated to a large degree the real investment decision-making portion of asset management − and that is the next frontier. And that is what Qraft is really focusing on enabling and accelerating.

AI Business: But isn't wealth management about the personal touch, a human being giving bespoke investment advice?

Nestor: AI, at its heart is about scaling, analysis, and data. And that data can come in many different forms, from traditional financial data to web traffic to a variety of different forms. Artificial intelligence allows the analysis and scaling of that at a speed that cannot be accomplished by human alone.

But, to get directly to the question, AI also can be very responsive to changing dynamics of the marketplace … in terms of the risk tolerance of any individual. In the last 30 years, … it almost didn't matter where you're invested (during) greatest bull market ever. But the last six to seven months had been a sobering reality that risk still exists, and markets can go down.

AI allows for and facilitates a very responsive strategy to those changing market dynamics, and those changing investor sentiments. We think that is one of the real areas of growth.

AI Business: How can AI help in risk management?

Nestor: AI has the ability to deal with, test repetitively and learn from the past and consistently adjust in its recommendation − and instructions are always overseen by humans.

There’s a couple of different ways that AI plays a role in that investment decision-making. AI can analyze tremendous amounts of data, to look at stocks and drivers of return, and build portfolios very much from the bottom up that reflect the most attractive securities based on those drivers’ returns, but also the dimensions of risks that any individual, or frankly, any asset manager, wealth advisor wants to accomplish. Its flexibility and scale are very powerful.

So if you want a portfolio that has the opportunity to add alpha (above-market gains) but have relatively tight tracking error or risk, then we can go for that and processes allow for that. Or if you want it to be highly tactical, to be more aggressive, to be more narrow in its approach and bottoms up that can be done.

One of the other dimensions is top down. We have proprietary capital markets models that look at the very wide level and sub-level, the different components of our world, and make very specific asset allocations, whether it be on rates, whether it be on credit, whether it be on equity dimensions, both domestically and globally, and make very specific recommendations as to an appropriate asset allocation in the current environment that we're in.

The third dimension is really analyzing and determining opportunities to improve on cost reduction and efficiency around the trading function. Specifically, looking at things like the depth and width of order book and what demand will be there on any individual stock or any individual security and improve the efficiency of returns. And it does that at a scale, scope, and speed that a human alone could never match. So it's really about man and machine and leveraging that human intuition.

AI Business: There's one problem with AI though that could be important for financial applications: the black box issue. AI cannot tell you why a certain investment was made − or is there a way to figure it out?

Nestor: It's definitely harder. But black box is … not a new label. It goes back to traditional quantitative practices that have been around for decades. However, there's this concept of explainability that cuts across not only artificial intelligence application in asset management, but other dimensionality and other industries. But it's particularly pronounced in certain aspects, because people want to understand why does the portfolio driven by AI hold Microsoft, not hold IBM?

It is admittedly very difficult to get at that level of precision, because they are almost always put together in a portfolio context. But there is also intuition that is built into that. We spend a lot of time talking with clients, what are the primary drivers at any point in time and what the model likes and dislikes, and we can give pretty decent indications of why the model likes something at a particular point in time. It's not absolute to translate it into weights because it's an iterative process. Risk reduction isn't always about the specific stock.

But the bottom line is we can give some sense of explainability and intuition, but it's not easy. And I will tell you, it's an evolving process that providers of artificial intelligence capabilities continue to need to focus on and do a better job of and that is definitely a focus of ours.

AI Business: Can you share some use cases with our listeners?

Nestor: Sure. One client … that is a platform was looking for distinguishing opportunities that are highly tactical and driven by artificial intelligence. They look at the world more through the lens of top down capital market opportunities and we are building for them a highly tactical capital markets model that will look to access over 45 different asset classes and sub-asset classes in a very responsive, multi-asset portfolio that is driven not just by purely alpha opportunities, but (also offers) a highly refined risk framework that overlays it − and we're going to do that in a very customized way.

That is a fairly typical ask from the clients that are engaged with us. They want to access the power of the AI, but they want to do it on their terms. And they have very specific asset class and risk dimensions that they want to accomplish, but they want to have AI be the fundamental driver.

Another area that we're working on is in the ESG world. We have a client that is interested in a more refined sub-theme with a particular passion on diversity and inclusion. We're … using one of the classic artificial intelligence techniques − natural language processing − to identify companies globally at great scale that have leading policies and procedures as it relates to inclusion and diversity and then building a portfolio from the bottoms up. … Artificial intelligence allows for that scaling in a very direct way.