The Advisor's Value After Information Gets Cheap

A client brings a retirement analysis to the meeting. An AI tool produced it in a few minutes. The scenarios look plausible. The prose is clean. The recommendations sound confident.

Then the real work starts. Which inputs are wrong? Which trade-off matters most? What else has to happen before anyone acts?

A thick report once signaled effort. That signal is fading. Advisors now have to show where their judgment changed the decision.

A polished report proves less than it used to

Wealth firms already use AI to search internal knowledge, summarize documents, analyze client feedback, and prepare communications. LSEG describes those uses across the industry. Vanguard draws a useful distinction between analytics-heavy work and empathy-heavy work, with greater automation potential in the first group.¹

Those industry accounts do not describe every firm, and they do not prove that an AI-assisted answer is correct. They do show how the cost of producing a credible-looking document is falling.

Polish can help a client understand the work. It cannot establish that the inputs are complete or the recommendation is suitable. Speed says little about whether someone noticed a fragile assumption, caught a conflict between goals, or understood the consequences of getting the call wrong.

This is where the value question gets uncomfortable. Firms have often made their effort visible through pages, projections, and meeting materials. When software can produce those artifacts quickly, the client needs another way to see what the advisor contributed.

The answer will vary by client. It may be a fact that changed the projection, a tax issue that changed the sequence, a family concern that changed the timing, or a risk that the client had not considered. The common thread is simple: the advisor's work has to alter something that matters.

People use AI before they trust it

Gallup's August 2026 study of financial guidance in the United States and Canada shows how use and confidence can pull apart. Among Americans who sought financial guidance in the prior year, 73% used internet research, 32% used a professional financial advisor, and 18% used an AI tool. In Canada, internet research also led, while 43% of guidance-seekers used a professional advisor.²

The confidence figures look different. Among all adults, 79% of Americans and 76% of Canadians reported at least some confidence in financial advisors' expertise. No more than three in ten in either country said the same about AI. Younger adults used AI more often, yet fewer than half of them expressed confidence in it.²

These numbers come from one cross-sectional survey. They cannot tell us whether confidence is rising or falling, and they do not follow the same people over time. They do show that trying a tool and trusting its judgment are separate choices.

That pattern makes sense. Someone might use AI to learn a term, organize a list of questions, or test a rough scenario. The same person may still want a professional to challenge the assumptions and take responsibility for the conversation that follows.

A March 2026 CFA Institute study found a similar split among more than 2,400 younger affluent investors across six markets, including the United States and Canada. About one-third had used generative AI for financial education. Human advisors remained their most trusted source of investment guidance.³ This is a narrow population, so it should not stand in for all investors. It does suggest that digital self-education and professional advice can coexist in the same client.

The fee question is still unsettled

Janus Henderson tested whether mentioning AI changed what investors said they would pay. Its 2026 survey included 1,000 U.S. investors age 25 or older with at least $250,000 in investable assets. Each participant saw one of four hypothetical profiles: a financial planner who used AI, a planner with no AI mention, an investment advisor who used AI, or an investment advisor with no AI mention. Participants chose an annual fee from 25 to 200 basis points.⁴

Average stated fee tolerance was 85 basis points for the planner described as using AI and 95 for the planner with no AI mention. The averages were 93 and 95 basis points for the two investment-advisor profiles. The report called the effect of mentioning AI modest.⁴

The experiment weakens the claim that clients will automatically discount an advisor who discloses AI use. It does not protect current fees. Each group contained 250 people, every profile was hypothetical, and stated willingness can differ from what a client accepts on an invoice or at renewal.

Another result needs the same care. Respondents who rarely or never used AI personally reported average fee tolerance of about 85 basis points. Frequent users reported about 113 basis points. Janus warned readers against treating that relationship as causal.⁵ Frequent users may be more comfortable with technology, more optimistic about it, or more willing to delegate. The survey cannot tell us which explanation matters.

So the fee evidence gives firms less reassurance than a quick headline might suggest. The presence of AI did not produce a large penalty in this experiment. Actual pricing, retention, and switching remain open questions.

Advisor value shows up in the decision

TransUnion's August 2026 survey offers another clue about what clients notice. Among 1,000 U.S. consumers with at least $20,000 in investable assets, 65% of current investors placed trust and reputation among their top factors when choosing a provider. Fees and pricing were selected by 49%.⁶

That comparison has limits. TransUnion sells identity, fraud, and data services. Its public release does not include the full questionnaire or weighting method. The result says nothing about whether a particular fee is fair, and it cannot prove that trust causes retention.

Still, the result fits a practical view of advice. Clients judge a relationship through the way information gets turned into a decision. An advisor may know that two accounts belong to one household, that a parent will need support, that a business sale is uncertain, or that a client is likely to abandon a plan during a sharp market move. Those details can change the recommendation before the final document appears.

Trade-offs matter too. A useful advisor makes the cost of each path visible: more liquidity can mean less long-term exposure; an earlier retirement date can leave less room for error; a simpler plan can sacrifice tax efficiency. The value lies in explaining which compromise the client is actually making.

Then comes coordination. A sound decision can fail when accounts, deadlines, family members, or outside professionals move out of sequence. Someone has to track the dependencies and notice when new information changes the plan.

Accountability ties the work together. The client should know who reviewed the evidence, where uncertainty remains, and who owns the recommendation. CFP Board emphasized human judgment, ethics, privacy, model risk, and governance in its August response to U.S. lawmakers.⁷ That is the position of one professional body, not a legal standard for every North American firm. It captures a basic expectation: a model output needs a responsible process around it.

A useful review starts with the client

A service inventory can tell clients what a firm offers. It rarely shows what changed because the firm was involved. Five questions can make that contribution easier to see:

1. What did the firm know about this client that changed the analysis?
2. Which trade-off became clearer during the conversation?
3. What people, accounts, or deadlines had to be coordinated?
4. Who reviewed the evidence and owned the final judgment?
5. What follow-through could the client observe after the meeting?

These are diagnostic questions. They do not set a fair price, score the quality of advice, or replace a firm's compliance process. Their purpose is to find the moments where the relationship affected a real choice.

The questions also help firms decide where disclosure matters. Janus asked respondents how they felt about several advisor uses of AI. Only 12% said they would be upset if an advisor used it to create educational content, and 13% said the same about administrative tasks. The share rose to 33% for investment recommendations and 40% for automatic texts or emails.⁴

Clients appear to care about the task, especially when judgment or personal communication is involved. That makes a generic statement such as "we use AI" much less useful than a clear explanation of where the tool enters the process, who checks its work, and who remains responsible. The survey measured stated comfort, so it cannot tell us whether those views improve advice or change client behavior.

Watch what clients do next

Most of the current evidence records attitudes. It tells us what selected investors say about confidence, provider choice, and hypothetical fees. It does not show how actual fees, switching, retention, or client outcomes will change as the tools improve.

There is a credible downside case. BCG's 2026 analysis describes scenarios in which AI causes substantial business-model disruption and, in a more extreme outcome, displaces advisor roles.⁸ BCG presents these as scenarios instead of forecasts. Firms still have to take the possibility seriously.

The next useful evidence will come from behavior: fee changes across comparable service models, renewal and switching data, repeated measures of confidence, outcome studies, and jurisdiction-specific rules for AI-assisted work. Those signals will say more than another survey about whether clients like the idea of AI.

For advisors, the immediate test is concrete. Can the client point to a fact the firm uncovered, a trade-off it clarified, a dependency it coordinated, or a judgment someone was willing to own? If the answer is yes, the relationship has a visible role in the decision. If the answer is no, a polished report will have to do too much of the explaining.

Footnotes

1. LSEG, "AI is redefining the wealth advisor experience", August 4, 2026; Vanguard, analysis of what AI can and cannot replace in financial advice, August 14, 2026. ↩

2. Gallup, "Where Americans and Canadians Turn for Financial Guidance", August 5, 2026. Associated Press reported that the U.S. portion surveyed 5,075 adults age 21 and older from March 20 to April 6, 2026, using a probability panel. ↩ ↩²

3. CFA Institute, "Next-Gen Investors: A Guide for Wealth Managers and Financial Advisers", March 24, 2026. ↩

4. Janus Henderson Investors with 8 Acre Perspective, "2026 Investor Survey: Perspectives on AI", Appendix C, accessed September 2, 2026. The report restricts reproduction; values are reported here for criticism and analysis, not as a reused figure. ↩ ↩² ↩³

5. Ben Rizzuto, Janus Henderson Investors, "Investor Survey: Does AI usage impact advisory fee sensitivity?", August 10, 2026. ↩

6. TransUnion, "Trust and Transparency Matter Most in Wealth Management Relationships", August 20, 2026. ↩

7. CFP Board, "CFP Board Highlights the Value of Human Advice as AI Rapidly Grows" and full RFI response, August 14, 2026. ↩

8. Boston Consulting Group, "AI and the New Economics of Wealth Management", May 27, 2026. ↩