Practifi’s Johnstone: Generative AI Is Best Used As A Development Tool, AI Won’t Replace Advisors Anytime Soon, And Innovation Must Be Balanced With Security
With wealthtech in constant flux as artificial intelligence revolutionizes our work and innovation presses breathlessly ahead, it’s easy for advisors and professionals to find one new service or product after another, without framing the holistic picture of what’s happening and where technology will take their practices, clients and livelihoods in the foreseeable future.
To filter out the noise and find firm direction, we spoke with Adrian Johnstone, CEO of Practifi, about key wealthtech issues facing the wealth management industry today. Johnstone walks us through how AI will shape the advisor space and how innovation always needs to be balanced with security concerns.
Different Flavors Of AI
The type of AI most recently making headlines worldwide is “generative” AI, which generates content. Combined with “natural language processing,” or the processing of language by machine learning, generative AI is able to draft written content that mirrors human language.
Johnstone points out that that AI comes in different “flavors,” not just the generative AI with natural language processing that has captivated the general public. “That leans you away from what people often think of as ‘advice tech,’ like portfolios, investment options and similar things that are very numbers-led, and not very natural language processing led. The generative AIs just aren’t well suited to that.”
“I’m not yet convinced that they’re really ready for gameday,” said Johnstone, noting that there could be problems with generative AI from a compliance and client experience perspective. “If I’m paying my advisor a significant fee, I don’t want to feel like an AI is responding to my questions.”
Well Suited For A Different Purpose
Johnstone points to a very different use for AI that wealthtech companies can take advantage of – reducing the cycle time to innovation. As clients feed ideas for product innovation to Johnstone and his firm, many of those ideas require a heavy resource use of developers, build cycles, test cycles and checking.
“Natural language processing is great for that. If I’ve got to code something, and a big part of it is repetitive foundational code, why should I pay a developer to write that when I can get the AI to write that part and the developer can do the hard bit at the end? The way we’re seeing it is: ‘How do we augment our engineering capability? How do we augment our quality assurance? How do we free developers from the repetition and minutiae and point them to the more challenging and rewarding work?’”
“When you reframe generative AI to be a development tool rather than trying to think about what investor experience you’re delivering with it, it becomes incredibly powerful, today, instead of having to think of how it may become powerful in the future.”
Can The SEC Get Comfortable With AI?
In the future, Johnstone sees a point when his firm adds generative AI capabilities to its portal, but at current AI is not performing at a level of accuracy or regulatory comfort to convince him to take that step.
“I don’t see the SEC turning around tomorrow in an industry that is still driven by wet signatures, to suddenly say they’re fine with AI. Will it get there? Yes, because ultimately the investor experience will win out in time. I feel like there’s a long road ahead there. Time will tell how comfortable the regulator can get with the role of these technologies.”
Innovation, Integration And Risk
Johnstone points out that at the same time that companies wish to move forward rapidly with integration, security risks continue to surface in areas such as data breaches.
While he strongly supports integration, “Not all systems are equal. Many legacy platforms, for instance, struggle with data encryption and the more contemporary forms of authentication. At Practifi, we don’t have that problem because we’re built on Salesforce.”
Johnstone says that many clients will ask him why his firm’s services don’t integrate with certain external systems, and his standard response is, “Because of your best interest.” As he explains, what may seem efficient can open a Pandora’s box of risk that advisors don’t need to understand, but do need to trust that their technology partners understand and don’t put the advisors in harm’s way. He points out that regulators put the onus on the firms to undertake due diligence on third party service providers.
“Part of the role of being a good citizen in a technology landscape is to be prepared to be the one who says, ‘We’re not doing this now.’ We have to put the security and protection layers in place, which is sometimes seen as stifling innovation.”
As an example, Johnstone points to the direction of data flow. Advisors often begin thinking they would like all custodial data to flow bidirectionally, but after learning that it means overwriting the custodian’s record based on any edit to the firm’s record, advisors hesitate and explain the nuances of what they actually need flowing into the custodian’s data and, conversely, what can’t be allowed. “In their eagerness for the outcome they forget the immense amount of detail that’s required.”
Where Is Advisor Technology Going?
Johnstone predicts a future that continues to focus on making “human advisors more efficient so they can invest more time in the client experience.”
Johnstone believes advisors should not worry about being replaced by AI in the near-term future. “It’s a tool for efficiency – it will always be that. Look at robo-advisors and the challenge that they had monetizing. Ultimately someone still wants to talk to a human. I think the advisor is very safe.”
He predicts that advice technology will either be used to move the need for human engagement further down the process to perform a validating role, in which the investor interacts with a person after engaging with tech-driven systems, or to deliver a much more curated experience to high net worth clients by feeding the advisor timelier information and prompting them on what they should do to support the client.
For the latter, Johnstone sees natural language AI systems starting to feed more information to the advisor rather than the client, and for both of these, he sees a focus on efficiency to build more scale without sacrificing experience.
Julius Buchanan, Editor in Chief at Wealth Solutions Report, can be reached at jbuchanan@wealthsolutionsreport.com.
1 comment
Comments are closed.