When AI Makes Financial Answers Abundant, What Becomes Even More Essential? | AI & FINANCIAL ADVISORS
Updated: Sep 3
Co-Authored: Clements Investment Management with R. LaRue, Social Research Specialist, 2026
INTRODUCTION: AI AND FINANCIAL ADVISORS
Artificial intelligence is rapidly changing one of the fundamental conditions under which financial advice has traditionally been delivered: Research that once took hours can increasingly be assembled in minutes.
Investment alternatives can be compared almost instantly. Complex financial concepts can be explained on demand. A sophisticated investor can ask an AI system to analyze a portfolio, compare strategies, model scenarios, interrogate assumptions, summarize an estate document or generate a remarkably plausible answer to a financial question before ever speaking with an adviser.
This is a meaningful advance. It may also produce unexpected consequences.
When sophisticated data retrieval, comparison, synthesis and first-pass analysis become available to nearly everyone, access to those capabilities becomes less scarce—and therefore less differentiating.
That raises a more useful question than whether artificial intelligence will replace financial advisers:
What becomes more valuable when data, analysis and plausible answers are available on demand?
For investors with relatively simple financial lives, technology may continue to eliminate work that once required professional assistance.
But wealth management rarely becomes simpler as life grows more complex and circumstances change.
The challenge at that point is seldom a shortage of information. It is determining which information matters, what is missing, which risks deserve attention, which opportunities actually fit, how changing or unexpected conditions alter the problem—and what a particular person or family should do when several individually rational choices lead in different directions.
That is the territory of accumulated experience and personal stewardship. As the analytical capabilities of AI become more powerful, current research points toward an important paradox: the distinctly human capacities to frame the right problem, discern what matters, integrate a client's financial and personal realities, anticipate consequences, and remain accountable for the counsel given may become more valuable—not less.
PART I: The Question Shapes The Answer – Why powerful analytic tools depend on how the question is framed…
The integration of artificial intelligence into professional wealth management is no longer hypothetical.
In June 2026, Rockefeller Capital Management announced a collaboration with Anthropic to build an AI-enabled wealth-management platform using Claude. The project is particularly instructive because Rockefeller serves high- and ultra-high-net-worth individuals and families—precisely the part of financial advice where client circumstances can become unusually complex.
The stated purpose of the technology is substantial: integrate AI into adviser workflows, deepen the insight available to advisers, synthesize complexity and improve the ability to deliver personalized service.
But equally revealing is what Rockefeller did not describe as obsolete.
Its announcement explicitly situated AI within a framework of judgment, context, long-term relationships, discretion, care and responsibility. The technology is intended to extend the capabilities of the adviser, while the adviser remains responsible for understanding the client and exercising judgment.
Much of the public conversation about AI and financial advice has been framed as a contest between machine capability and human intelligence. That framing risks obscuring a more consequential distinction.
AI is extraordinarily well-suited to many information-intensive and computational parts of financial work. It can search a much larger field than a person could reasonably search, compare alternatives, detect patterns, summarize documents, generate scenarios, surface questions and test assumptions. Properly used, those capabilities can materially enlarge the field available to both advisers and clients.
A competent adviser in the emerging environment should therefore be technologically capable rather than technologically defensive.
But greater analytical capability does not settle what may be the most important question. Before asking what the answer is, someone has to determine whether the problem has been framed correctly in the first place.
An AI-generated answer does not arise independently of the inquiry that produced it. How a question is framed, what information is supplied, what assumptions are embedded, what context is omitted and what the user fails to ask can materially shape the response.
This matters because clients cannot reliably supply context they do not realize is relevant, challenge assumptions they do not recognize they are making, or deliberately investigate risks that experience has not yet taught them to see.
The quality of the inquiry therefore becomes part of the quality of the answer.
A seasoned adviser contributes before the analysis begins: identifying the real problem, discovering missing context, challenging a preferred conclusion, recognizing conflicts between stated goals and observed behavior, and knowing when tax, estate, legal, insurance or other expertise must enter the conversation.
AI can be remarkably capable at answering the question it is given. Expertise begins one step earlier: knowing whether it is the right question.
What should an investor do when some of the variables most likely to change the answer may remain invisible until an experienced adviser knows to ask about them?
An investor approaching retirement after forty years of accumulating assets may technically have the capacity to tolerate a particular level of risk while having little psychological capacity to watch a lifetime of savings fluctuate once a paycheck disappears.
An inheritor may suddenly possess substantial assets without having developed the experience, confidence or financial identity that usually accompanies the gradual creation of wealth.
A business owner preparing for a liquidity event may appear to be making an investment decision when the deeper decision concerns identity, family, purpose, taxation, control and the shape of the next thirty years.
Two households with comparable balance sheets can therefore require very different advice.
The more consequential the wealth, the more frequently financial decisions become life decisions expressed in numbers.
Artificial intelligence can contribute greatly to analyzing those numbers. Whether it can independently understand the life around them is a very different question.
PART II. Where Affluent Investors Still Draw the Line—What affluent and high-net-worth investors are revealing about AI, trust and consequential decisions
Investors appear willing to use AI extensively for exploration and analysis. However, their behavior changes as they approach decisions with greater personal and financial consequence.
What should an investor do when some of the variables most likely to change the answer may remain invisible until an experienced adviser knows to ask about them?
Investor behavior is beginning to reveal that distinction.
HSBC’s 2026 global research of nearly 10,000 affluent and high-net-worth investors found that 73% were already using AI for financial and investment purposes. Yet only 12% identified AI as the most influential factor in their most recent investment decision. Financial professionals and institutions remained the principal source of investment ideas for 62% of respondents.
The U.S. results make the divide even clearer.
Fifty-seven percent of affluent U.S. investors reported using AI for financial and investment tasks. They use it for analysis and research, strategic support and as a second opinion. Yet only 7% said AI was the most influential factor in their last investment decision, while 59% said their last investment idea came from financial professionals or institutions.
Among U.S. high-net-worth investors with at least $2 million in investable assets, 67% identified financial professionals and institutions as the leading source of an investment idea, compared with 16% who cited AI.
This is not evidence of resistance to technology. Quite the opposite, it suggests investors are beginning to create their own division of labor: AI is being used to explore; to investigate; to become better informed. And then, many investors will seek strategic expertise, context, and a human professional being prepared to stand behind a recommendation.
Separate 2026 research from Janus Henderson points in the same direction. Its study of 1,000 U.S. affluent and high-net-worth investors found broad acceptance of advisers using AI for educational and administrative purposes. Comfort dropped substantially as AI moved closer to personal communication and actual investment recommendations. Investors also expressed concerns involving bias, data security and trust in AI-generated recommendations.
What investors appear to be preserving is not simply a preference for “human touch.” It is the expectation that consequential advice will be interpreted through their particular circumstances by someone able to question the analysis, exercise independent judgment and stand behind the recommendation.
The pattern is important because it challenges two simplistic predictions at once.
Investors are neither rejecting AI nor surrendering financial decision-making to it.
They are learning where it helps.
And they are beginning to reveal where they still want a person.
PART III. The Questions Before a Correct Answer—Why accumulated experience, context, and personalized client stewardship matter when wealth becomes real life.
For years, discussions about financial technology have repeatedly returned to some version of the same question:
Will technology replace the financial adviser?
AI makes this question more immediate, but not necessarily useful.
A better question may be: what an adviser is for when financial information, computational analysis and increasingly sophisticated first-pass recommendations become inexpensive and ubiquitous—and when the quality of the result can still depend materially on the quality of the inquiry.
If the principal value of an adviser were simply possessing information that the client could not obtain elsewhere, AI would represent a profound threat to that value proposition.
But sophisticated wealth management has never been only an information or analysis business. At its best, it is a business of correct and detailed problem-framing, complex interdependent discernment, experienced judgement about human behavior surrounding financial decisions.
This distinction becomes particularly visible as wealth grows—or as life changes.
Consider the questions that tend to arise around a major inheritance, retirement, business sale or other significant transition: How much risk is appropriate now that accumulation is becoming distribution? Should a concentrated holding be reduced despite its tax consequences—or retained despite its risk?
How should competing goals among a spouse, adult children and future generations be balanced?
What should remain liquid?
When is an attractive investment nevertheless inappropriate?
How should an estate strategy interact with investment strategy?
What happens if longevity, health, inflation, markets or family circumstances unfold very differently than expected?
How much wealth is enough to support one’s own life, and when does the purpose of the capital begin to change?
AI may improve the inputs into these decisions considerably—and may help surface possibilities a human adviser might otherwise miss.
But the proliferation of data, analyses and plausible alternatives can make another skill more important: distinguishing what applies, what is missing, what is being assumed, and what deserves weight in this particular life.
That changes where the experienced adviser may add the greatest value.
The adviser is progressively less valuable as a gatekeeper to information and more valuable as a framer of the problem; a curator and interpreter of competing evidence; a challenger of faulty assumptions and confirmation-seeking; a translator among investment, tax, estate and life consequences; a behavioral counterweight at precisely the moments when intelligent people can make poor decisions; and, in the case of a fiduciary adviser, a professional accountable for placing the client’s interests at the center of the recommendation. (This latter distinction being essential.)
Long experience adds another dimension: pattern recognition grounded in actual people and their consequences.
Twenty or thirty years of advising people through market cycles, family transitions, retirements, deaths, inheritances, business exits, exuberance, panic and unexpected change does not merely produce a larger storehouse of financial facts. It can develop a disciplined ability to see omissions and consequences before they are obvious to the person living through them.
Over time, that experience can produce informed judgement about people as well as markets.
A seasoned adviser may know that a client who says “I’m comfortable with the risk” has behaved very differently during every significant decline.
The adviser may know that a financially optimal estate decision will create exactly the family conflict the client most wants to avoid.
The adviser may recognize when the apparent investment question is actually a question about security after the death of a spouse, fear about retirement, discomfort with newly inherited wealth or reluctance to relinquish the identity attached to a family business.
And a genuinely personalized advisory relationship accumulates something else over time: context.
Not simply age, assets, objectives and a questionnaire-based risk score, but the history of how this particular person makes decisions, what has frightened them before, what they value, where spouses differ, what obligations they feel toward children, what they regret, what “enough” means to them, and which financial risks would have the greatest human consequences if they materialized.
Technology can store information about a client.
The harder question is whether having the information is equivalent to knowing the client—and whether a model can recognize which unspoken, omitted or changing facts are the ones that should alter the question.
And that kind of judgment is not created at the moment a question is asked. It accumulates—through years of seeing people make decisions, live through their consequences, change their minds, encounter the unexpected and discover that what matters to them financially is often inseparable from what matters to them personally.
For wealth management in the AI era, that distinction may prove decisive.
As sophisticated analysis and plausible answers become increasingly available on demand, the scarce resource may no longer be the answer itself.
It may be a trusted person with enough experience to know which answer belongs to you.
PART IV. THE VALUE THAT ACCUMULATES—What decades of personalized stewardship allow an adviser to see, ask and anticipate that a new relationship—or a stand-alone analytical tool—may not
Research can help identify the capabilities of artificial intelligence and the continuing importance of context, judgment and trust. What research cannot supply is the accumulated experience of sitting across from real people while consequential financial decisions become consequential life decisions.
That experience matters because the question a client initially brings into the room is not always the question that ultimately needs to be answered.
Someone may arrive asking whether to sell an investment, retire at a particular age, accept a certain level of risk, make a large purchase, retain a concentrated holding or alter an estate strategy. Each may be a legitimate financial question. But an experienced adviser may notice that the decision turns on something the client has not yet included in the analysis—a spouse’s differing sense of security, an unspoken family obligation, prior behavior during market stress, a health or longevity assumption, tax consequences elsewhere in the plan, or simply a fear or aspiration the numbers alone cannot reveal.
One client came to me with what sounded like a narrow question about the Affordable Care Act. She did not believe she had enough income to afford another health plan. In the same conversation, she mentioned that she had canceled a trip to Paris after a friend backed out because she did not think she could afford to travel alone.
As we talked further, it became clear that affordability was not really the problem. She had substantial investment assets, but she carried a deep fear of overspending and eventually “dying broke.” There were also aspects of her portfolio and borrowing that did not fit what she actually needed.
After several meetings and a comprehensive financial plan, she began to see that her resources gave her considerably more freedom than she believed. As her confidence grew, she began traveling and allowing her wealth to support more of the life she had worked to create.
An AI system could have answered her ACA question correctly. The value of the advisory relationship was recognizing that the ACA was not really the question.
The importance of that distinction is easy to miss. The analytical answer to the original question might have been perfectly reasonable. The problem is that a correct answer to an incomplete question can still lead someone in the wrong direction.
Experience also changes what an adviser is able to recognize over time.
A new analytical system—or even a new adviser—can be given a detailed record of assets, liabilities, income, objectives, family structure and stated preferences.
A long relationship contains a different kind of information. It includes remembered behavior: how a person actually responded when markets fell; which risks sounded tolerable in theory but felt very different in practice; what changed after retirement, an inheritance, a death, a marriage or a business transition; where spouses see security differently; which decisions repeatedly produce anxiety; and which stated financial goals are really serving something deeper.
In my experience, one of the hardest transitions is moving from a dependable paycheck to relying on income from a portfolio. Even when the financial plan is sound, it can take many clients months to truly trust that the income is sustainable and that they can continue living the life they planned for.
I have often seen another version of this after the death of a spouse, particularly with surviving spouses who were less involved in the household finances. Along with the personal loss comes the sudden responsibility of understanding how their financial life will work from that point forward. The questions may be similar whether the relationship with me is new or longstanding. The crucial difference is that an existing client does not also have to build trust with a new adviser while everything else in life is changing.
I also regularly encounter investors whose portfolios carry more risk than their financial lives actually require. The issue is not simply age or a risk-tolerance score. The portfolio may never have been evaluated against the income the client needs, the lifestyle the assets are meant to support, or how that person actually responds when markets fall. Risk makes sense only in the context of the larger financial life.
The distinction is subtle but important. A database can preserve facts. A relationship can accumulate meaning around those facts.
And sometimes the value of experienced counsel appears not in finding a more sophisticated answer, but in resisting an answer that appears attractive.
A mathematically efficient strategy may be wrong for a particular family. A tax-efficient choice may produce consequences elsewhere that matter more. An investment a client strongly wants may fit the portfolio on paper but conflict with a larger financial objective. A decision that is ultimately appropriate may still be inappropriate today.
The adviser in these situations is not merely validating a client's preference. The responsibility sometimes requires challenging it.
A longtime client began showing signs that his decision-making was changing. During a semiannual review, I learned that he was considering significant changes to his estate plan involving a family business. The building would go to one child and the operating business to another—even though the siblings were estranged.
Because I knew the family history and had developed a longstanding working relationship with the client's trust attorney, we were able to slow the decision down and help the client reconsider the proposed changes. What mattered in that moment was not simply what the estate documents said. It was what years of knowing the client and his family allowed us to recognize about the consequences those changes were likely to create.
The decisive information was not hidden in a financial document. It existed in the history of the relationship.
That ability to challenge a conclusion becomes especially important as technology makes it easier to obtain analysis supporting almost any line of inquiry. A sophisticated investor can ask more questions, examine more alternatives and arrive better informed than at any previous point in the history of wealth management.
But being better informed does not eliminate the possibility of wanting the wrong thing, asking the wrong question or giving too much weight to an attractive answer.
Experienced stewardship also extends beyond the investment portfolio itself.
A decision involving investments may create tax implications. An estate decision may alter investment needs. A liquidity event may change income requirements, estate planning, family expectations and the psychological meaning of wealth simultaneously. Knowing when the problem has exceeded the boundaries of investment management—and when an attorney, accountant, insurance professional or other specialist needs to enter the conversation—is itself part of judgment.
Deciding whether income should come from an IRA or a taxable investment account is rarely an isolated investment decision. Tax consequences, Social Security, required distributions and the client's larger income plan can interact in ways that change the answer. When evaluating distribution or Roth conversion strategies, we coordinate with the client's CPA and estate-planning attorney so that one technically sound decision does not inadvertently create a problem somewhere else.
This is also where accountability enters the picture. A recommendation is not simply an analytical output delivered at one point in time. Life continues after the recommendation is made. Markets move. Health changes. Families change. Tax laws change. A spouse dies. An inheritance arrives. A business is sold. Something that was appropriate five years earlier may become wrong without having been wrong when it was first chosen.
The adviser who has been part of that history remains in the relationship as the consequences unfold.
Any recommendation I make begins with research, but the research is only one part of the decision. I then have to test the recommendation against the client's Financial GPS—their income needs, risks, objectives, changing circumstances and the life those assets are meant to support.
Unless something material changes in the client's life, goals, planning assumptions or the investment case itself, part of my responsibility is helping the client avoid abandoning a sound long-term plan because of short-term noise or fear. In an ongoing advisory relationship, the responsibility does not end when a recommendation is made. It includes continuing to monitor, revisit and adapt the advice as circumstances change.
That continuing responsibility may be one of the least quantifiable elements of financial advice and one of the most consequential.
Artificial intelligence will continue to improve. It will almost certainly become better at searching, synthesizing, modeling, comparing and identifying possibilities. Good advisers should make intelligent use of those capabilities.
The more interesting question is what happens to the value of human judgment as those capabilities become abundant.
The answer may not lie in defending work that technology can increasingly do well. It may lie in becoming clearer about the work that was never primarily informational in the first place: asking what has not been asked, understanding the human consequences around the numbers, recognizing patterns accumulated across years, challenging a conclusion when necessary, coordinating complexity and remaining accountable after the recommendation has been made.
After all these years, I think what clients most need from me when a consequential decision arrives is not more information. They need to see the real options clearly, understand the tradeoffs, and have someone who knows them well enough to help recognize which choice actually fits the life they are trying to live.
I don't see AI as something an adviser should defend against. If it can help me research faster, test assumptions, compare possibilities or see something I might otherwise miss, I want to use it. But the output still has to be judged against what I know about the client, their Financial GPS and the life the recommendation is actually going to affect.
As financial answers become increasingly abundant, those capacities do not appear less relevant.
They may become the scarce asset.
Co-Authored
Matt P. Clements, Managing Principal; Clements Investment Management
36 years advising clients
Ronda L. LaRue, M.S, Social Research Specialist. 40 years corporate research & individualized mentorship
ANNOTATED RESEARCH SOURCES
The following sources inform the research and observations presented in this article. They document current developments in AI-assisted wealth management, characteristics of large-language-model decision support, and the attitudes and behavior of affluent and high-net-worth investors. They should not be interpreted as establishing that human advisers, artificial intelligence, or any particular combination of the two necessarily produces superior investment outcomes.
Rockefeller Capital Management. (2026, June 10). Rockefeller Capital Management Collaborates with Anthropic to Build AI-Enabled Wealth Management Platform.https://www.rockco.com/media/rockefeller-capital-management-collaborates-with-anthropic-to-build-ai-enabled-wealth-management-platform/
Razavi, A., Soltangheis, M., Arabzadeh, N., Salamat, S., Zihayat, M., & Bagheri, E. (2025). “Benchmarking Prompt Sensitivity in Large Language Models.” Advances in Information Retrieval: ECIR 2025, Lecture Notes in Computer Science, 15574, 303–313. Springer.https://doi.org/10.1007/978-3-031-88714-7_29
Ranaldi, L., & Pucci, G. (2023). “When Large Language Models Contradict Humans? Large Language Models’ Sycophantic Behaviour.”https://arxiv.org/abs/2311.09410
Vaccaro, M., Almaatouq, A., & Malone, T. (2024). “When Combinations of Humans and AI Are Useful: A Systematic Review and Meta-analysis.” Nature Human Behaviour, 8, 2293–2303.https://doi.org/10.1038/s41562-024-02024-1
HSBC Holdings plc. (2026, June 24). The Trust Threshold: AI Makes Investors Bolder, but They Want Human Judgement to Make Decisions.https://www.hsbc.com/news-and-views/news/media-releases/2026/the-trust-threshold-ai-makes-investors-bolder-but-they-want-human-judgement-to-make-decisions
HSBC Bank USA. (2026, June 25). The Trust Threshold: The Majority of U.S. Investors Use AI to Explore, But Only 7% Say It Drove Their Last Major Decision.https://www.about.us.hsbc.com/newsroom/press-releases/the-trust-threshold-the-majority-of-us-investors-use-ai-to-explore
Janus Henderson Investors. (2026, May 19). 2026 Investor Survey: Perspectives on AI.https://www.janushenderson.com/corporate/article/2026-investor-survey-perspectives-on-ai/
