Key points are emerging as high-net-worth investors increasingly turn to AI chatbots like Claude and ChatGPT for portfolio allocation and tax planning advice. Top wealth management firm executives highlight that AI can help clients engage more targeted, in-depth discussions with advisors, yet it also poses significant risks. While AI chatbots push advisors to enhance their competitiveness, the human touch remains irreplaceable, especially during market turmoil and crises.

Many wealthy clients, capable of hiring top-tier financial advisors, now use AI tools like Claude for asset allocation strategies and tax planning references, according to several wealth management executives. Matthew Fleissig, co-founder and CEO of Pathstone, an asset management firm overseeing $185 billion, stated, "In my view, ChatGPT is now the world's largest financial advisor. It's a licensed investment advisory firm managing $185 billion." Industry professionals believe that clients seeking multiple perspectives through AI can formulate more sophisticated questions, improving the quality of advisor interactions. However, this approach carries risks, as AI can provide incorrect advice and potentially leak personal financial data. Michael Zeiner, managing partner of WE Family Offices, a family office, noted, "For highly complex financial issues, average clients struggle to discern whether AI is making factual errors, fabricating content, or offering viable solutions."

Sean Dunlop, an analyst at Morningstar, suggests that as large language models evolve and become more widespread, the wealth management industry—particularly firms targeting mass affluent clients—must deliver more value to justify fees. AI is squeezing industry margins, pressuring stock prices in related wealth management sectors. Examples include Altruist launching an AI tax planning tool in February, and OpenAI introducing personal finance features in May. Dunlop believes AI won't completely replace traditional advisors but will drive industry upgrades, boosting advisor efficiency. "The real trend lies between two extremes: overall service standards rise; fewer advisors serve the same asset scale, with a slight expansion in serviceable assets; and some clients who previously relied on advisors now manage their own assets," said the Morningstar equity research director. "At minimum, AI raises the industry baseline. If advisors still let clients leave half their retirement account funds idle as cash, AI sounds an alarm."

AI is deeply integrating into client financial processes. Pamela Lucina, chief fiduciary officer and head of trust advisory at Northern Trust, noted that about 18 months ago, she first observed clients using AI to verify financial plans, a trend now more common. "Many clients explicitly tell us they'll run questions through AI first," Lucina said. "Their core request isn't always a standard answer, but confirmation that we can provide detailed evidence and past case studies." Potential clients also use large models to vet advisory firms. Previously, only billionaires would request formal financial plans, but Lucina estimates that among clients with assets as low as $100 million, half now proactively seek plans; many admit to using ChatGPT to compile detailed question lists. Lucina sees this as saving time and enhancing meeting efficiency. "Discussions no longer waste time on basic education, but focus more on client goals and tailored family paths. Public information isn't a threat; it deepens conversations." Zeiner reported that five months ago, clients routinely mentioned using large models to review asset allocations. His firm even considers running all financial advice through Microsoft Copilot first, helping advisors anticipate AI-related client questions. He views large models as a client's second pair of eyes, largely beneficial. "AI encourages clients to learn, participate in financial decisions, and seek multiple perspectives. Whether clients question us or verify our plans, it's deep engagement."

AI's risks and shortcomings are significant. Zeiner worries most about clients failing to spot AI errors. For instance, a client uploads trust documents to a large model, and days later, the chatbot fabricates document details. He also witnessed basic mistakes: ChatGPT told a client that two ETFs were identical, but one tracked the same index with equal weighting, while the other was market-cap weighted—clear structural differences. Lucina personally tested AI on capital gains tax strategies, and the model made complete calculation errors. "I often use AI to find logical flaws and counter-check, where it's useful. But for practical financial advice, AI's error rate is high."

A critical risk is privacy leakage. Even with paid personal AI accounts, uploading financial data poses exposure threats. Regulated financial firms purchase enterprise-grade AI with strict data protection protocols. Zeiner explained that WE Family Offices signed with Microsoft Copilot, ensuring all entered data isn't used to train public AI. "We educate clients on secure AI use based on their high-net-worth status and privacy needs," Zeiner said. "If AI records all meetings, where is chat history stored? Is it secure? Who has access? These must be clarified."

Human service remains irreplaceable. Fleissig noted that Claude and ChatGPT excel at data analysis but cannot deliver actionable, customized financial plans. Large language models are semi-objective tools, struggling with highly subjective, personalized decisions. Experienced advisors integrate multiple factors to help clients choose plans suited to their families and life goals. "The human relationship in financial services has unique value, involving nuanced, tailored considerations." AI can argue for investing in unlisted Anthropic but can't secure private fundraising access for clients. "Investment access relies on personal networks and long-term advisor experience. Quality investment resources still require human connection, which AI can't facilitate." Vince Lumia, a 27-year Morgan Stanley veteran overseeing over 16,000 advisors, noted that in bull markets, human service value is often overlooked, but during market turmoil and uncertainty, clients heavily depend on human advisors. "When markets are strong, everyone thinks investing is easy. But one certainty is that markets fluctuate; they don't only rise, and disruptions occur. Whenever crises hit, high-net-worth clients prefer seeking advice from real people."