Most advisors agree that AI technology improves efficiency, including streamlining workflows and communicating more effectively with clients. But the jury is out on whether these budding but expensive investments are worth it.
“This technology is very expensive. I mean, some of these technologies cost hundreds of thousands of dollars a year.” [AI] “We've seen a lot of changes in the past few years,” said James Bogart, CEO and president of McLean, Virginia-based Bogart Wealth, which works with AI vendors and builds advanced capabilities in-house. Significantly improve value creation and efficiency. ”
Recent
This comes especially after the U.S. Securities and Exchange Commission (SEC) cited two advisory firms on March 18 for proudly claiming to use AI-based investment decisions when making investment decisions. This incident comes at a time when people are already worried about the explosive spread of AI. It wasn't.
read more:
SEC Chairman Gary Gensler said, “Time and again, the emergence of new technology can generate buzz from investors and false claims by those purporting to use the new technology.'' “I've seen it,” he said in a news release about the fraud case. “Investment advisors should not mislead the public by saying they are using AI models when in fact they are not. This kind of AI laundering hurts investors.”
The SEC expanded its warning to companies that partner with AI vendors.
How to measure AI technology is difficult, but options exist
Many AI vendors specializing in asset management are privately held companies, often startups, and therefore don't report financials like publicly traded companies, making it difficult to know their underlying profitability. .
It is also difficult to determine the accuracy of the AI in the tools being sold. One reason for this is that AI programs are still learning and are prone to so-called hallucinations and erroneous outputs, as seen in ChatGPT, a widely used AI language model.
read more:
One metric used primarily by AI developers and testers is called the F1 score, which measures the precision and recall score of a machine learning model. Basically, it calculates the percentage of correct responses relative to the amount of information received.
“So if you want to receive a response on 100 data points and 94% are completely accurate, your F1 score is effectively 94%,” Co-Founder and Chief Product Officer says Danny Lohfink. “So it's a combination of volume and accuracy,” he tells Wealth.com, a Phoenix, Ariz.-based technology-based platform for estate planning for financial advisors.
Lohrfink said that when Wealth.com first started developing AI technology within its trust business, the company's F1 score was 71%.
“Everything is now about 90% more accurate,” he said, adding that it took about 18 months to teach the AI model. “We've been able to train and tune the model, and we've collected more data, and it's gotten better and better. That's why the scores keep going up.”
Advisors want AI to improve efficiency, but at what cost?
When it comes to AI providers, pricing structures vary widely, with some advisors charging through a license or subscription-based model, and others charging based on the number of users or assets they manage.
Pricing is also evolving, with some companies such as SalesForce moving from licensing to a cost-per-data usage model for cloud services.
“As we move into this AI world, we are starting to move more towards allowing customers to offer pricing models that take into account their usage tiers and take into account what they consume from their data cloud. ” said Michelle Feinstein. , general manager and vice president of global financial services at Salesforce, a cloud-based customer relationship management software provider in San Francisco.
Feinstein and Bogart agreed that the price is worth paying if AI providers can demonstrate how their models create efficiencies in workflow and customer experience.
The Back Diamond Wealth Platform, for example, “has been a great partnership, and it's definitely my most expensive piece of technology,” Bogart said. “But the truth is, they provide a tremendous amount of value.”
Bogart said that as a company with nearly $3 billion in assets under management, the number one question when considering purchasing an AI vendor is: “Can a company my size handle it?”
Whether a vendor builds the AI technology or an advisor in-house, there will be costs. And the larger the company or the amount of data input needed to train the AI, the more expensive these platforms will be to develop.
“It's like building a house, right? I don't want to know the proposal or the budget.” [set] “And all of a sudden, you didn't know what you were doing, so we doubled or tripled that amount,” he said, “I've never had an experience like that. Because it was there.”
