Evergreen.ai study: advised AI users act unchecked at nearly triple unadvised rate

Investors who already pay for financial advice are the most likely to take a chatbot's word on money matters without checking it, according to the Evergreen.ai Financial Decisions Study reported by InvestmentNews. Among respondents who work with an advisor, 63.8% said they had asked an AI tool a question about their money in the past 12 months, versus 38.5% of those without one. Among advised AI users, 41.9% said they acted on an answer without checking it against another source or a professional, nearly triple the 14.4% rate among unadvised AI users.

The study analyzed 1,981 responses from U.S. adults aged 25 to 75 who make or share their household's financial decisions. Across the full sample, 48.4% had asked a chatbot about their finances. Of AI users who answered a follow-up question, 77.8% acted on what they were told, and 28.7% did so without verifying. Those last two figures describe a different population than the advised-versus-unadvised split.

The report's authors caution the data shows an association, not proof that advice drives the behavior. AI users as a group skew wealthier and are more likely to have an advisor, which the report says explains why they were less likely to report making a major decision without knowing its tax impact: 15.4% versus 41.6% of nonusers. That gap, per the report, reflects who the users are more than what the technology did for them. The authors suggest people used to seeking professional guidance may treat a chatbot as one more adviser in the room — a hypothesis the study does not test.

Evergreen.ai is a service of Evergreen Wealth Advisors, the Miami-based registered investment adviser launched by PayPal and Intuit cofounder Bill Harris. "The financial advice gap in this country is real, and Americans are desperately seeking financial guidance," said Harris, the firm's founder and CEO. "They are looking for answers wherever they can find them, and nearly half of those we surveyed have already turned to AI. The question is whether the tools they are using were built to deserve their trust." InvestmentNews reported that M1 launched its own AI advisor aimed at households turned away by wealth managers. Benchmarking by advisor fintech Saturn, cited in the report, found the best-performing free model still returned incorrect or incomplete responses 56% of the time.

For an RIA, the practical question is what the firm can see. A client who asks a chatbot about a Roth conversion and acts on the answer produces nothing the advisor can review; a client who asks, is asked to verify, and gets a written response produces a record. That is why the 41.9% deserves more attention than the 63.8% adoption figure: adoption builds a habit, and the habit is what a client file has to survive. Evergreen is selling the opposite wager — that a tool can answer well enough to be trusted. The 56% error rate the report cites is the part of that wager still open.