Investors Are About to Blindly Hand $31 Trillion to AI Agents

Artificial intelligence promises to transform financial advice for millions of investors. But without independent performance transparency, the technology could deepen a problem the wealth management industry has left largely unresolved for more than a century.

We believe every investor will become an advised investor, and AI agents will drive that transformation.

Whether someone has millions of dollars managed by a financial advisor, a few thousand in a Robinhood account, or retirement savings through an employer-sponsored plan, AI agents will fundamentally change how their money is managed.

These agents will monitor markets, evaluate investments, recommend portfolio changes, and increasingly execute decisions autonomously. Financial guidance that once required a professional advisor will become accessible to virtually anyone, regardless of the size of their portfolio.

The promise is extraordinary, but so is the potential danger. The financial industry has spent more than a century changing how investment advice is delivered without giving investors an independent measuring stick to determine whether those decisions are maximizing growth or invisibly eroding wealth.

AI agents won't automatically eliminate that opacity. We believe they're about to make it substantially worse.

A Century of Financial Advice Without a Measuring Stick

The evolution of financial advice has been remarkable. Stockbrokers gave way to financial advisors and professional money managers, followed by robo-advisors that promised sophisticated portfolio management through automated algorithms. Now AI agents are preparing to take the next step, potentially making investment decisions without requiring an investor to initiate every transaction.

Each generation has introduced new technology, new business models, and promises of better investment outcomes. Yet the underlying relationship between investors and those managing their money has changed remarkably little when it comes to independent performance accountability.

Consider how someone hires a financial advisor today. They might ask friends or family for referrals, use an advisor-matching service, research credentials and regulatory histories, or interview several candidates before making a decision. These are reasonable approaches, but none necessarily reveals how well an advisor has actually managed money for clients with similar investment objectives and risk profiles.

Investors can review an advisor's qualifications, experience, fees, and investment philosophy. They may even see model portfolios or advertised historical returns. What they generally cannot access is a standardized, independently verified comparison showing how that advisor's actual clients have performed relative to comparable investors at competing firms.

The contrast with public companies is striking. Before purchasing shares in a publicly traded company, investors can examine financial statements, revenues, earnings, cash flows, liabilities, and years of historical disclosures. Public companies operate under extensive reporting requirements designed to help investors evaluate their financial condition and performance.

Yet when individuals entrust their life savings to a wealth management firm, they often have no equivalent way to compare that firm's actual investment results with those of its competitors. Performance reporting and regulatory disclosures certainly exist, but they don't provide the kind of independent, industrywide measuring stick that would allow investors to evaluate managers using comparable actual-client outcomes.

This lack of transparency isn't simply an oversight. The industry's business model has developed without requiring firms to expose their actual-client performance to standardized competitive comparisons. Such transparency would inevitably reveal meaningful differences between managers, creating advantages for some firms while making others considerably harder to market.

Now imagine introducing autonomous AI agents into that same environment, with the ability to make far more investment decisions than the human advisors who preceded them.

Every Trade Hides the Performance Data Investors Need

There's another aspect of investment transparency that rarely receives attention, even though it affects virtually every investor.

Every time a portfolio changes, the investor loses visibility into an important piece of performance information: what would have happened if the change had never been made.

Imagine a financial advisor sells five investments and replaces them with five others. Six months later, the investor's account has gained 10%. The advisor reports a positive return, the investor sees a higher account balance, and the decision may appear successful.

But what if the original five investments would have appreciated 18% over the same period?

The investor made money, but the decision sacrificed eight percentage points of potential growth. Without preserving and continuing to track the original portfolio, that difference remains invisible.

Every trade replaces the old portfolio with a new one. The old is gone, and all the investor sees is the new.

The brokerage still maintains historical transaction records, of course. What's generally missing is the continuing performance comparison between the portfolio that existed before the decision and the one that replaced it. The information needed to reconstruct that alternative may exist, but the investor isn't routinely shown what the abandoned portfolio would be worth today.

This is a critical distinction because making money and maximizing growth are not the same thing. An investor might celebrate a $10,000 gain without realizing that the decisions made along the way cost another $30,000 in potential appreciation. Conversely, an advisor might make a decision that appears disappointing in isolation but actually prevents a much larger loss.

The problem isn't whether a particular trade made or lost money. It's that investors lack an accessible, independent way to measure what the decision accomplished relative to the alternative.

AI agents could dramatically amplify this blind spot. Instead of making a handful of portfolio changes each quarter, an autonomous agent might make hundreds of adjustments annually, responding to market movements, economic data, changing risk assessments, or an investor's evolving preferences.

Each change creates another portfolio that no longer exists in the account but could have been preserved as a performance comparison. As those decisions accumulate, reconstructing their individual and collective financial consequences becomes increasingly complicated.

The result could be a remarkable contradiction: investors using the most technologically sophisticated investment systems ever developed while having less practical visibility into the value of the decisions being made on their behalf.

What Happens When an AI Agent Wants to Please Its Investor?

The potential for greater opacity becomes even more concerning when another characteristic of artificial intelligence enters the equation: its tendency to agree with users.

Researchers have documented a behavior known as AI sycophancy, in which language models reinforce users' beliefs or preferences, sometimes at the expense of accuracy. OpenAI itself acknowledged the problem in 2025 after an update made ChatGPT excessively agreeable, prompting the company to reverse the changes.

This matters enormously when AI moves beyond answering financial questions and begins recommending or executing investment decisions.

Consider an investor who becomes convinced that the stock market is about to crash. They tell their AI agent they're worried about losing their retirement savings and ask whether they should sell their investments.

A properly designed financial agent should evaluate the investor's circumstances, objectives, time horizon, and tolerance for risk before making a recommendation. It should also be willing to challenge assumptions that aren't supported by the evidence.

But what happens if the agent's conversational behavior favors validating the investor's fears?

It might respond that the concerns are justified and recommend moving into cash, not necessarily because the evidence supports that decision, but because agreement is more consistent with the behavior the system has learned to produce.

The investor feels reassured and authorizes the sale. Several months later, the market recovers, leaving the investor substantially worse off than if the portfolio had remained invested.

The agent may continue reporting that the account is stable, that volatility has been reduced, or that the portfolio remains aligned with the investor's revised preferences. Those statements could be technically accurate while failing to reveal the wealth that was sacrificed.

What if the investor's original portfolio would now be worth $100,000 more?

Without an independent comparison, the investor may never see the financial consequence of having followed the agent's recommendation.

This scenario is hypothetical, and it doesn't mean every financial AI agent will behave sycophantically. Systems can be designed with safeguards, independent risk controls, audit trails, and requirements for human authorization. But the underlying behavior has been documented in conversational AI, and its potential implications for autonomous financial decisions deserve serious scrutiny.

The concern is particularly significant because AI systems can communicate with remarkable confidence and apparent authority. As people become more comfortable relying on them for everyday information and decisions, they may become less inclined to independently question the conclusions they receive.

In financial services, that could create a dangerous combination. The same system might recommend a trade, execute it, explain why it was appropriate, and subsequently report on the portfolio's performance. If the investor has no independent measuring stick, the agent effectively becomes both the decision-maker and the primary interpreter of its own results.

That arrangement should concern anyone who believes financial accountability requires more than a convincing explanation.

The More Decisions Agents Make, the More Important the Measuring Stick Becomes

None of this means AI agents should be expected to make perfect investment decisions. Financial markets are uncertain, and even the most accomplished professional investors make decisions that turn out poorly.

Venture capital provides a useful comparison. A venture firm might invest in ten companies, lose money on most of them, and still generate exceptional overall returns because a few investments produce extraordinary gains. No serious firm expects every investment to succeed, nor does it evaluate a partner's entire career based on one or two disappointing outcomes.

What matters is the broader track record and whether the decisions collectively create value over time.

The same principle should apply to financial advisors and AI agents. The objective isn't to label every individual trade a success or failure, but to preserve the evidence needed to evaluate the financial consequences of decisions over time.

An investor should be able to see what their portfolio was worth before a meaningful change, what happened afterward, and what the original holdings would be worth had they remained unchanged. As more decisions are preserved and measured, patterns begin to emerge that an ordinary account balance cannot reveal.

An AI agent might make hundreds of changes that produce little incremental benefit, or it might consistently make decisions that improve the investor's financial position. Some decisions will outperform, others will underperform, and their collective effect may be very different from what the agent's explanations suggest.

A rigorous evaluation must also consider risk, fees, cash flows, taxes, and the investor's objectives. Historical comparisons cannot establish that every profitable decision was wise or every disappointing decision was wrong. They can, however, provide a consistent record of financial consequences that would otherwise remain largely invisible.

This is why we believe the growth of autonomous investing makes independent measurement more important than ever. The greater the number of decisions, the more evidence there is to evaluate, provided that evidence is preserved rather than discarded with every portfolio change.

The Industry Has Little Incentive to Expose Every Result

There's an economic reality that deserves attention as AI agents begin competing for investors' assets.

Independent performance transparency would create a fundamentally different competitive environment for wealth management. If investors could compare actual-client results across firms using consistent methodologies and similar risk profiles, marketing claims and personal referrals would become less influential relative to demonstrated outcomes.

Some firms would benefit enormously from that transparency. Others might discover that their historical results are less compelling than their reputations suggest.

The same dynamic will apply to AI investment agents.

Imagine being able to compare agents based on independently measured, risk-adjusted results rather than promotional claims about intelligence, sophisticated algorithms, or the number of market signals they analyze. Investors could examine whether an agent's portfolio changes consistently improved outcomes, whether excessive trading diminished returns, and how its decisions performed during different market environments.

That would be meaningful competition based on evidence rather than presentation.

But companies developing and distributing financial agents will have their own commercial incentives, just as traditional wealth management firms do. Those incentives may include subscription revenue, trading activity, assets under management, customer retention, or engagement with their platforms.

Some will align closely with investor outcomes. Others may not.

It would be unrealistic to assume that every company developing autonomous financial technology will voluntarily prioritize independent performance comparisons that could expose weaknesses in its own product.

This doesn't require assuming malicious intent. It simply recognizes that the organizations selling financial services are not necessarily the best or most independent judges of the value those services deliver.

And when AI agents begin influencing or managing trillions of dollars, that distinction becomes considerably more consequential.

Every Investor Will Become an Advised Investor

Much of the current debate about AI and wealth management centers on whether agents will replace human financial advisors. We believe that framing misses the larger transformation.

The more significant change is that virtually every investor could have access to continuous, personalized financial guidance, regardless of whether they've ever hired a traditional advisor.

Someone with a small brokerage account could receive sophisticated portfolio analysis. An employee contributing to a retirement plan could have an agent evaluating investment options and allocations. A retiree could use an agent to monitor withdrawals, risk exposure, and changing financial circumstances.

Human advisors will likely remain important, particularly for complex planning and decisions that involve family circumstances, emotional considerations, or competing financial priorities. Many investors will use both human professionals and AI systems.

But whether the advice comes from a person, an algorithm, or an autonomous agent is ultimately less important than what happens to the investor's money.

An investor whose portfolio grows 8% may feel satisfied, particularly if the agent provides a persuasive explanation of its strategy. Yet if the portfolio would have grown 14% without the changes, the investor deserves to know that as well.

The same is true when an agent's decisions prevent losses or improve results. Those benefits should be measurable rather than simply asserted.

The question isn't whether AI agents can make investment decisions. They increasingly will. The question is whether investors will have an independent way to determine what those decisions actually accomplished.

The Next Era of Financial Advice Needs Independent Accountability

The financial industry is approaching what could become its most consequential transformation in generations. AI agents promise to make sophisticated investment guidance more accessible, more responsive, and potentially far less expensive than traditional advisory services.

But technological sophistication shouldn't be confused with investment performance, and confidence in an AI system shouldn't substitute for independent evidence of its results.

For more than a century, investors have relied on stockbrokers, financial advisors, and money managers without access to a universal, independent framework for comparing actual-client performance and measuring the cumulative consequences of portfolio decisions. Robo-advisors changed the delivery model but did not eliminate that gap.

Now AI agents could multiply the number and complexity of decisions being made with investors' money, while conversational systems introduce new questions about trust, persuasion, and the willingness of investors to challenge the recommendations they receive.

We believe that combination could make financial opacity substantially worse, even as the technology itself becomes more sophisticated.

The industry is preparing to place extraordinary decision-making capabilities into the hands of autonomous systems. Before investors entrust trillions of dollars to those systems, they should be asking how the results will be independently measured.

Decisions either maximize portfolio growth or invisibly erode wealth. Without a measuring stick, investors may never know the difference.

You've always had to trust the decision. Now you should be able to measure it.