Will AI Replace Human Financial Advisors? Evidence from My New Working Paper

I am pleased to share my current working paper, Silicon Counsel: Profiling the AI-Advice Seeker and the Substitution Effect in Retail Financial Planning, now available on SSRN.

The rise of artificial intelligence in consumer finance has triggered a major question for academics, practitioners, and regulators alike: will AI-generated financial advice replace human financial advisors, or will it mostly work alongside them? This paper tackles that question using nationally representative U.S. survey data and offers a more nuanced answer than the simple “AI replaces humans” narrative that often dominates public discussion.

Why this question matters

AI is already reshaping the financial-services landscape. Consumers increasingly encounter algorithmic tools in investing, budgeting, payments, and platform-based financial management. Yet there is still limited evidence on who actually wants AI advice and, even more importantly, whether that interest reflects substitution for human advice or complementarity with it.

That distinction matters. If AI mainly substitutes for traditional advisors, then the implications for advisory firms, labor markets, and regulation are profound. If, instead, AI mainly complements existing human relationships, then the future of financial advice may be hybrid rather than fully automated.

What the paper does

Using the 2024 National Financial Capability Study, merged with the companion Investor Survey, I examine interest in AI-generated financial advice across a large and nationally representative sample of U.S. adults. The analysis combines several empirical approaches, including binary probit models, multinomial logit specifications, propensity score matching, and a Heckman two-stage selection model.

This design allows me to move beyond a simple yes-or-no view of AI adoption and instead distinguish between different advisory preferences and investor profiles.

Main findings

First, the typical person interested in AI advice is not financially naive. The evidence suggests that the AI-advice seeker is more accurately described as digitally engaged, relatively risk-tolerant, and institutionally trusting. In fact, FinTech engagement emerges as the single strongest predictor of interest in AI-generated financial advice, stronger than financial literacy, trust, or overconfidence taken separately.

Second, interest in AI advice is not monolithic. Among investors, the paper identifies two especially important groups:

• Complement Seekers: individuals who want AI advice alongside human financial advice. These respondents tend to be more trusting, more overconfident, and more likely to hold larger portfolios.
• AI Substitute Seekers: individuals who want AI advice instead of human advisors. These respondents tend to be more objectively knowledgeable, more sceptical of traditional institutions, digitally fluent, and more likely to hold relatively smaller portfolios.

Third, the overall relationship between human and AI advice appears more complementary than substitutive. After adjusting for observable differences between advisor users and non-users, investors with experience using a human financial advisor remain more likely to express interest in AI advice. In other words, exposure to human advice seems to increase openness to AI rather than crowd it out.

What this means

One of the most important takeaways from the paper is that the future of financial advice is unlikely to be captured by a single, dramatic replacement story. The evidence points instead to a segmented market.

For many consumers, especially highly engaged investors, AI is likely to become an additional layer of support inside an already existing advisory relationship. In that sense, AI may help advisors scale information processing, portfolio analysis, and routine guidance, while human professionals continue to provide judgment, reassurance, behavioral coaching, and relationship-based trust.

At the same time, the paper also shows that there is a real substitution-oriented segment in the market. These consumers are not simply passive or uninformed. Quite the opposite, they are often knowledgeable, digitally comfortable, and unconvinced that traditional advisory channels offer enough value relative to cost. That finding should be taken seriously by incumbent advisory firms.

Implications for regulators, firms, and educators

For regulators, the results suggest that existing frameworks may need to adapt to a world in which human and algorithmic advice coexist. Questions of disclosure, responsibility, suitability, transparency, and performance monitoring become even more important when consumers rely on both channels at once, or when purely algorithmic platforms serve investors independently.

For traditional advisory firms, the findings are partly reassuring and partly cautionary. They are reassuring because the evidence does not point to immediate wholesale displacement of human advisors. They are cautionary because a meaningful AI-substitution segment already exists, particularly among smaller but sophisticated investors for whom low-cost algorithmic solutions may be attractive.

For consumer advocates and financial educators, the findings highlight the growing importance of digital financial literacy, not only traditional financial literacy. As AI tools become more prominent in finance, unequal access to digital confidence and FinTech familiarity may widen existing gaps in the use of high-quality financial guidance.

A note on limitations

Like all early work on emerging technologies, this study has limitations. The survey captures expressed interest in AI advice rather than actual adoption, and the data are cross-sectional rather than longitudinal. The results therefore offer strong evidence on patterns of demand and heterogeneity, but they should not be read as the final word on how consumers will behave once AI advisory tools become even more widespread and more sophisticated.

That said, the paper provides an important first step by offering a nationally representative profile of the AI-advice seeker and by showing that the substitution-versus-complement question cannot be answered well with a simple binary framework.

Read the paper

You can access the full working paper here:

Read the full paper on SSRN

I would be very glad to hear comments, suggestions, and feedback from colleagues, researchers, finance professionals, and readers interested in the future of AI in retail financial planning.