Advisors shouldn’t take their eyes off of robo-advisors and digital advice.

Though advisors have become less concerned about robo-advisors disrupting the traditional wealth management industry, new artificial intelligence applications within digital advice platforms present a direct threat to advisors’ businesses, said Daniel Faggella, head of research at Emerj Artificial Intelligence Research, a firm that conducts market research on AI for enterprise clients.

“When we talk to vendors who have raised $50 million to $100 million, they feel like robo-advisors are just starting to gobble the market,” said Faggella. “A robo-advisor, even with AI, is not going to be able to make me feel warm and comfortable in the same way that sitting in front of somebody does. We’re 15 years away from technology being able to do that.

“But in reality, that kind of personal touch is only important to two categories of people: people with very high levels of needs—what I would call the high end of the market—and for lack of a better term, old people.”

Younger generations at all wealth levels, however, are very comfortable managing their money using technological tools, said Faggella. “Young people are just going to put their money somewhere that doesn’t charge them a crazy fee. They don’t want the fees and complexity that mom and dad were dealing with. As the younger generations replace the older ones, there’s more money in younger hands, and the pressure on financial advisors increases. I think we’re going to see a big shift to robo-advisors in the next five years.”

According to Faggella, the only advisors who are “safe” are ones who work with clients at the highest levels of wealth and complexity—the rest of the industry teeters on the verge of disruption.

Stories about AI’s power to transform the wealth management industry may be premature, he said, but the technology will likely first have widespread impact in a handful of specific areas.

One of these areas ripe for AI infusion is regulation and compliance, he said, because firms are willing to open up their wallets to solve compliance issues or avoid them altogether.

“One thing AI might be able to identify effectively is the actual trading activity that we think might violate some kind of rule, something that sends a red flag up. That pattern can be identified,” he said. “Another side of this is phone and text transcripts. AI can comb through broker and advisor conversations with people to determine if there’s anyone potentially leaning into breaking the rules.”

According to Faggella, five times more money is going into AI-driven compliance applications as is going to marketing, sales and customer service combined.

Similarly, a lot of funding and research is going into cybersecurity applications of AI, where it can be used to identify risky behavior before it’s exploited by criminals.

As AI becomes more commonplace in regulatory compliance, fraud detection and analysis jobs will disappear, said Faggella.

AI will also be gathering and analyzing more novel types of data than it did in the past, he said—analyzing the sentiments of different market participants, or using satellite analysis of retailers, ports and other venues of commerce and trade. That will offer an edge to users.

He said any task having to do with data entry and paperwork is likely to be automated.

Though it will significantly disrupt the insurance and banking industries, AI is less likely to have an immediate impact on back-office employees in wealth management, said Faggella, as many of the most easily automated functions have already been outsourced to third parties or technology.

“The obnoxiously boring paperwork processes are going away; those roles will be gobbled up,” he said. “We no longer need people actively, manually pecking into systems. Some of these people can be used for higher level tasks.”

But when more back-office functions at financial advisory firms are replaced with AI and other technologies, there are going to be permanent job losses, he said, as most back office personnel lack the characteristics to become great client-facing advisors. “If you work in a back office, you might not be enough of a glad-handing person to be successful as an advisor and maintain a book of business. There’s a personality factor there.”

The customer service applications of AI, things like chat bots, get a lot of press but they aren’t receiving a lot of research and development funding right now, he said. That’s because firms are less likely to discuss the compliance and security enhancements they are making in the public eye, but will gladly talk about the relatively small amount of money spent to develop a client chat program.

“We call it the lens of incentive—a company will let people know about its technology initiatives when it believes it will make the company look better in front of who matters, like journalists,” said Faggella. “Is Wells Fargo ever going to say that they just spent $10 million with Darktrace, an AI cybersecurity vendor, to protect the data of their North American customers? Probably not. It creates an additional security risk, and it scares people. So financial services companies tend to talk most about how the customer service experience is going to be transformed.”

He downplayed the influence and value of most innovations in marketing technology, noting that many of the early client- and advisor-facing artificial intelligence applications are now defunct, suffering from a lack of financial support and low rates of adoption.

“Tiddlywinks money is going into marketing applications in financial services,” he said. “Most of these apps are bloviated garbage; there’s really not much going on there in terms of investment.”

There are two major obstacles to adopting AI into anything customer-facing, said Faggella. The first is that the immediate money that financial institutions are willing to spend on next-generation technology tends to focus on risk-related issues. The second is that there are only so many software developers in the world with the skill set to develop the customer-facing applications of AI, and they tend to be focused in Silicon Valley.

A tertiary reason that AI is unlikely to find a place in advisors’ front-office operations is the risk of losing clients.

“If I have an internal discovery application at my firm that helps employees retrieve files faster, nobody will quit if it breaks 20% of the time, but a chat bot that breaks 5% of the time—nobody is going to use it,” he said.