How an early adopting advisor assesses AI tools

Advisors also need to be critical of AI outputs. Fact checking and clear sourcing are key to any AI-generated content, and challenging the AI model on anything it has generated can often catch those AI errors. While AI can often tell users that their ideas, questions, and perspectives are all “great,” Zagari says that the models can be trained to give clearer and more critical responses to their human users. AI can be trained to be less sycophantic, making it more effective. Zagari also notes that sometimes AI output can appear impressive, but it doesn’t clarify. Every communication that he uses AI to edit or draft, he says, must then be tested under the simple question of whether it makes things more or less clear for the client. AI generated content can seem impressive, but clarity should always be the goal.

Zagari applies the same tests to AI tools that vendors will try to offer him as well. For operational AI tools, Zagari notes that decisions tend to happen at a dealer level. Those decisions, though, should be made with a view to how any AI tool integrates with the other tools on a dealer’s shelf, he says. Many AI vendors will focus on marketing solutions for advisors and when he deals with those vendors, Zagari says he avoids AI lead generation tools “like the plague.” As much as he uses AI in his communication, he doesn’t trust these marketing platforms to make a positive first impression. They give the impression that his practice is impersonal, he wants it to be anything but. As much as he sees AI benefitting his practice, he argues that authenticity should remain core to every advisor’s approach.

“Be authentic, be real. Use AI as an enhancement, not as a replacement in your practice,” Zagari says.