The traditional path into financial advice is being quietly dismantled by artificial intelligence. The entry-level analyst and paraplanner roles—data gathering, report assembly, follow-up documentation—that once served as the apprenticeship for new advisors are increasingly handled by algorithms. This shift comes at a precarious time: McKinsey & Company projected in February 2025 that the industry could face a shortage of roughly 100,000 advisors by 2034, as retirements outpace new entrants and client demand surges.
The question is no longer whether AI will absorb more of the analytical workload—that is already happening. The real challenge is whether the next generation of advisors will still develop the judgment, communication skills, and client intuition that no machine can replicate. Four practitioners are tackling this head-on, each with a distinct strategy.
Hands-on first, AI second
Joseph Conroy, founder of Harford Retirement Planners in Bel Air, Maryland, an independent practice affiliated with Private Advisor Group, compares his approach to a medical residency. New advisors shadow experienced colleagues, sit in on client meetings, take notes, and draft follow-up summaries and emails from scratch. "We pretend AI doesn't exist in the early learning stages of a new advisor," Conroy said. "They 'do things by hand' and once they master it, then AI can start to reclaim their time back."
The rationale is not nostalgia for manual work. It is about understanding the reasoning behind a recommendation before delegating its assembly. AI has eliminated much of the need for those foundational tasks, Conroy acknowledges, but they remain the vehicle through which new advisors learn the "why" behind workflows and communication style. One task he will not let new advisors hand to AI before demonstrating proficiency: client email communications. Trust is the currency of the advisory business, and clients can detect a copy-pasted language-model message. "We do not want to compromise the value in our recommendations with the distracting thought of 'did my advisor send this or did they just get lazy and have AI type this email?'" Conroy said.
The art of the second question
Chris McMahon, founder and CEO of MFA Wealth and Aquinas Wealth Advisors, sees a specific trap in relying on AI before a new advisor has developed interviewing instincts. AI makes recommendations based solely on the hard data presented to it. A client who says they are comfortable leaving a surviving spouse with $50,000 per year will receive a plan built around that number—unless the advisor probes further. "An experienced advisor would be able to ask more probing questions and hold a mirror up to the client," McMahon said. In his example, the right follow-up—noting that the couple spent $23,000 on travel last year and had set a goal of visiting one new state per year with their family—typically produces a very different answer. The client realizes the original figure does not reflect their actual goals, and the income replacement number moves to $100,000. Only then, McMahon says, should AI come in to build the precision plan around those goals.
McMahon believes the traditional three-to-five-year analyst track is not disappearing but compressing. He now structures junior advisors' days so half is spent collaborating with seasoned advisors and half on analytical work. The result, in his view, is that the client-facing phase of a young advisor's career can begin in year two rather than year five. "Without the teaching, mostly observing these interviews with experienced advisors, this skill set won't develop, and the next generation of planners will fall short of their mandate of helping people achieve their true goals," he said.
Moving training upstream
Devon Klumb, director of sales at Betterment Advisor Solutions, the New York City-based custodian and technology platform for independent RIAs, frames the debate differently: AI automating the assembly of analysis is not new. The profession has been offloading technical construction to software—amortization tables, tax projections, financial planning models—for decades. What is new is losing the incidental connection to the numbers that used to come with doing that work. "Not plugging in the inputs is different from not knowing them," Klumb said. "Even if a new advisor never builds the analysis from scratch, they still need to review the inputs and question the outputs."
His training approach moves the developmental work upstream rather than eliminating it. New advisors at his firm are placed in front of clients earlier than the traditional track, given a defined role in the meeting, and held accountable for assembling their own recommendations before those recommendations reach a client. The debrief then focuses on the client's reaction—what they hesitated on, what they moved past quickly—rather than the technical accuracy of the plan. The task Klumb will not allow AI to touch until a new advisor has demonstrated they can do it unassisted is the discovery conversation: sitting with a client without AI tools and working out what that client actually wants their money to accomplish. "What tells me if they are ready to do it without help is whether they can explain the resulting recommendation in plain language, without notes, and articulate why it fits that particular client," Klumb said. "If they cannot, more tooling only makes them worse."
The broader industry is watching these experiments closely. As a recent student survey showed, AI anxiety is already deterring some from entering the field, contributing to a net loss of 4,000 advisors in 2025. Meanwhile, some advisors are finding new revenue streams by training AI systems that could eventually automate their own roles. The challenge for firms is to ensure that the next generation of advisors is not just a byproduct of automation, but a deliberate product of thoughtful training.
