- EquBot / Art Amador
- Equbot’s AI Powered Equity ETF uses IBM’s Watson technology to construct a stock portfolio, employing machine learning to make rational investment decisions.
- The original idea for the fund was synthesized in a classroom at UC-Berkeley, where founders Chida Khatua and Art Amador met during an entrepreneurship class.
- Khatua’s background in AI and machine learning complemented Amador’s history in private wealth management, and the duo decided to launch an exchange-traded fund.
When Art Amador worked in private wealth management at Fidelity, his clients expected him to know absolutely everything.
Whether it related to global markets, macroeconomic factors, specific companies, or full sectors, their curiosities were wide ranging – and Amador wondered if he’d ever find a way to be the all-knowing oracle they desired.
That all changed one day in the fall of 2014 when Amador was pursuing his MBA at the Haas School of Business at the University of California at Berkeley. As part of an entrepreneurship class, he was placed in the same cohort as a long-time Intel engineer and machine-learning specialist named Chida Khatua, and the two got to talking. That conversation led to what its creators say is the world’s first AI-powered exchange-traded fund, one built on technology that could change the paradigm for how computers are used to invest.
The fund – powered by IBM’s Watson supercomputing technology – didn’t end up launching for a few more years, but its roots can be traced back to that fateful first conversation at Berkeley.
“I was telling him it was impossible to have infinite knowledge about every stock, and about everything going on in markets,” he tells Business Insider. “I told him that there’s simply too much information out there and not enough time to distill it into actionable ideas.”
As it turned out, Khatua had been researching for years how to sift through massive amounts of data in a way that extended far beyond human capabilities. With two master’s degrees in computer engineering – including one from Stanford – he worked at Intel for 18 years, mostly focusing on machine learning.
“His background – in artificial intelligence and machine learning – was the perfect use case,” Amador says. “We started talking about how that could apply to the equity markets.”
Birth of an ETF
Even though the early groundwork had been laid for what would eventually become their newest venture, Khatua and Amador went their separate ways after the program ended. But the gears in Khatua’s head never stopped turning, and in September 2016 he invited Amador to join him in building a product that would combine their respective areas of expertise.
Amador took some time to think about it. In his mind, the result would be an AI-powered quantitative hedge fund, and he wasn’t sure if he wanted to give up his job at Fidelity for that. But Khatua had other ideas: He wanted to build and launch an ETF.
To him, the ideal application for his technology was to get it into as many hands as possible. And if he combined it with Amador’s investment prowess, they could build an ETF available to be traded by the average person with a brokerage account.
“Working at Intel gave me insight into how machine learnings and AI technology is maturing and how the benefits it offers can really be maximized,” Khatua tells Business Insider. “It gave me a unique perspective, and I asked myself for a while when the right time would be to go out and create some product that can help many people.”
Acting like a rational investor
A big part of Amador’s decision to ultimately join Khatua in pursuing an ETF was the latter’s acceptance into the highest tier of the IBM Global Entrepreneurship Program. After all, his machine learning and AI efforts were powered by the company’s Watson supercomputer.
That gave Khatua $125,000 with which to pursue his idea, and it provided Amador crucial validation for the endeavor. He joined up shortly thereafter, and the duo launched Equbot.
Then they put Watson to work. The eventual result was the recently launched AI Powered Equity ETF (ticker: AIEQ), which analyzes more data than humanly possible, all in the pursuit of building the perfect portfolio of 30 to 70 stocks. And the technology enables it to do that while constantly analyzing information for 6,000 US-listed companies.
But there’s a wrinkle. Equbot’s AI model is built to act like a rational investor. In addition to analyzing regulatory filings, quarterly news releases, articles, social-media postings, and management teams, it’s also designed to assess market sentiment and weed out potentially faulty inputs – including so-called fake news.
“A rational investor looks at a company as a whole and they draw insight into what’s right looking at the complete picture,” Khatua says. “The AI model helps us do that. The technology doesn’t only help you decide what to do; it can also educate you on why it’s happening.”
The technology doesn’t only help you decide what to do; it can also educate you on why it’s happening.
That’s a key element of AIEQ and one that sets it apart from the hedge funds that use AI to construct trading strategies. Khatua says many of those models function as a “conceptual black box,” because the presence of certain stocks can’t be explained in a rational way. In his mind, Equbot’s ETF offers the best of both worlds: It’s based on a mountain of analysis and the stock-picking methodology can be explained.
“We know why something’s in our portfolio after our system chooses it,” Amador says. “‘The system picked it’ is not usually an explanation that investors will buy.”
Further, the machine-learning aspect of AIEQ is crucial in avoiding human error. Amador points out that even if a firm had 6,000 analysts each responsible for reading 150 to 200 articles about one stock each day, that work would have to be cross-referenced against the findings of all other employees, then funneled into one objective opinion.
“Humans don’t have the speed, capacity, or retention to do this,” he says.
The story so far
AIEQ has slid 0.9% since its launch on October 18, while the benchmark S&P 500 has risen 1.6%. The biggest laggards in the fund are Lifepoint Health, Newell Brands and Vista Outdoor, which have each dropped more than 20% over the period.
But it’s far too early to judge the success of AIEQ based on five weeks of returns. The more telling statistic is the volume of shares traded. The ETF has seen an average of 259,000 units change hands daily, a strong showing for a fledgling fund. It had about $70 million in assets on Monday, roughly 10 times its size during the first week of trading.
The way that Khatua and Amador see it, interest in their product will continue to grow as long as personal bias continues to cloud investment decisions – something they see happening even at the highest level of professional money management.
“You can remove that by making this investment process more autonomous, as we’ve done,” Amador says. “It’s nothing against people. It’s just human instinct.”