SMITH BRAIN TRUST – Are you ready to entrust your retirement savings to a robot? Investment advice is increasingly coming to clients from nonhuman sources, whether they realize it or not, as more money-management firms look to automated services to augment their team of human advisers and deepen the science behind stock-picking. Some firms are adopting hybrid services that couple algorithmic services with human advisers, creating a less-actively managed fund, often with a lower fee structure.
But, at the same time, there is also a more deliberate move toward so-called robo-advisers, with some clients entrusting their savings to fully automated investment tools, replacing human advisers altogether. Alberto Rossi, assistant professor of finance at the University of Maryland's Robert H. Smith School of Business, has been studying the rise of robo-advice and its role in the financial industry.
He says the rise is driven mainly by two factors: the increased computational power and the ever-more-abundant data that are now available to private companies. These days, every transaction and customer decision is recorded, and that results in "tremendous amounts of data" to analyze, he says.
"Conventional statistical techniques are not always suited to data-rich environments," Rossi says, "while machine-learning algorithms are designed to excel exactly in these applications."
With his latest research, Rossi and his co-authors hope to plumb how individuals make financial decisions. "In particular, we want to understand whether a fully robotized system, where a virtual adviser gives the customer advice is as effective as one where the advice is delivered by an adviser that explains the suggestions of the robot," he says.
Anecdotal evidence suggests that investors are comfortable following the guidance of robo-advisers when the stakes are low, but as their wealth exceeds the $100,000 mark, they start to seek out the "human touch," he says.
Like many industries, financial institutions are looking to artificial intelligence as a potential means to serve as many clients as possible at the lowest cost possible. "Moving forward, the simpler services – such as concierges at hotel – will probably be replaced by apps and software," Rossi says, citing OpenTable's restaurant reservation service as an example. But for services that involve a high level of trust and customization, such as financial advice and medical advice, the digital transformation is a steeper hurdle In those industries, a hybrid model makes sense.
"At one end of the spectrum, we will have apps like OpenTable that help you book restaurants and give you suggestions on what restaurant to try next. For these services, human presence will likely disappear," Rossi says.
"At the other end of the spectrum, we will have companies like Redfin that will try to match real estate buyers and sellers," he continues. "While using machine-learning algorithms to help real estate brokers, these companies will rely on the ‘human touch,' because the decision to purchase a house is a very difficult one."
Meanwhile, the medical sector may also benefit from machine-learning algorithms, with doctors themselves using them to help diagnose patients. Still, Rossi adds, "the human interaction is going to be important in this context." Human beings are fallible. And because of that, some people will always resist entrusting a person with their money. Others will resist the opposite, looking for a sort of George Bailey "It's a Wonderful Life" relationship on which to base their investment decisions.
That's why investment firms need to get the hybrid "exactly right," Rossi says. "They will need to exploit machine-learning algorithms to prevent advisers from giving poor advice," he says. "At the same time, they will need to employ humans to answer the potentially complicated questions and concerns that customers constantly have."
The firms that strike the right balance between the two, he says, are the ones that are likely to succeed.
GET SMITH BRAIN TRUST DELIVERED
TO YOUR INBOX EVERY WEEK