Recasting retirement: 3 ways AI is changing client-adviser conversations

A few years before retirement expert Oscar Vives was born, Hollywood debuted an enduring classic that predicted the rise of artificial intelligence.

Luckily, no personal financial advisers were harmed in the making of the film.

“The Terminator, back in the ’80s, they predicted the robots taking over, and we are now living in that time period,” Vives, CPA/PFS, CFP, said. “So the future is now.”

Vives and fellow adviser David Blain didn’t predict that AI would lead to the demise of their profession during their June session at ENGAGE, but they did discuss how AI is changing the client dynamic.

They highlighted three ways that clients are injecting AI into the retirement planning puzzle but explained why advisers still hold the key to truly solving the equation.

“AI accelerates the thinking,” Blain said, “but the adviser has to own the advice.”

Clients arrive armed with AI-generated questions

Vives presented four slides full of questions — 40 in all — that prospective clients might ask after prompting AI to help them hire a retirement adviser.

“Clients,” Vives said, “are going to show up very prepared.”

Are you a fiduciary 100% of the time? What is the total all-in cost I should expect each year? How do you help clients manage market downturns during retirement?

If the prospect of playing 20 questions twice over with a prospective client isn’t appealing, don’t fret: Good questions allow a good adviser the opportunity to showcase their expertise.

Plus, advisers can make sure they’re prepared by reviewing Vives’ list of questions or by asking AI to help them organize their thoughts in advance of a meeting.

Clients use AI for a second opinion

Vives recently heard this from a client.

“Man, I’ve been working with Claude and it’s been amazing. It’s blowing my mind. By the way, I uploaded all the holdings in the portfolio that you guys manage for me, just to see what Claude has to say about it.”

The client then shared that Claude questioned the strength of some holdings within a 401(k) and suggested that a fund in the taxable portfolio was redundant.

Vives explained to the client that the minuscule 401(k) holdings were the only options available in the 401(k) and that the potentially redundant holdings had embedded capital gains that meant selling wasn’t a wise choice.

“It’s interesting,” Vives said, “because it allowed us to show our value proposition. But, be aware that clients are doing that.”

Clients ask AI to draft an entire retirement plan

While some clients are using AI to check up on an existing plan, some are taking it a significant step further, according to Vives.

“They’re saying, ‘Hey, I created my retirement income plan with ChatGPT. Can you review it? And you don’t even really have to do that much work because I’ve already done all the work.’”

That prompted Vives and Blain to create a case study, feeding a fictional couple’s financial details into ChatGPT and then asking a specific set of questions related to the couple’s potential path to retirement.

Notably, the initial ChatGPT response varied when Vives entered the prompt and when Blain did — with only Vives receiving a set of clarifying questions from AI. Similarly, a client asking AI about retirement on their own may get varying degrees of detail in response but may not be able to recognize that AI’s answer likely is incomplete.

“At least at this moment,” Vives said, “there’s not cohesive coordination amongst the robots.”

Vives said AI was correct and consistent about some big-picture items, such as whether the couple could retire at age 61 or whether the higher earner should delay Social Security until age 70.

However, the responses often lacked the necessary nuance that an adviser brings to the table. For example, just one of three AI responses offered annual cash flow projections for the retirement years, and Vives and Blain presented a laundry list of crucial considerations that AI entirely left out, including health care planning before Medicare eligibility and inherited IRA strategies.

Blain detailed how advisers can use their expertise in concert with AI’s computing power to supercharge support of clients well beyond what clients with a keyboard can do for themselves.

“Without adding skill, you’ve just got general-purpose AI. With skill, that’s where we can add value,” Blain said. “How we use AI to our benefit is applying our knowledge and expertise to it.”

Editor’s note: If you didn’t attend ENGAGE, you still can access this session. Those who purchased an all-access pass to ENGAGE can view this and other archived sessions.

— To comment on this article or to suggest an idea for another article, contact Bryan Strickland at Bryan.Strickland@aicpa-cima.com.