AI Is Expanding Wealth Management, Not Replacing Advisors

Almost every industry is seeing professionals concerned that AI could take away their jobs, especially in white-collar, knowledge-driven sectors.

This is, of course, true of programming, writing, and other tasks where modern AI based on LLMs (large Language Models) is especially proficient. But this is also true for finance professionals, who are increasingly seeing AI and so-called robo-advisors as potential competitors to traditional human-made wealth management.

A new study by two Chinese researchers at Shanghai University challenges this idea. Analyzing the effects of LLMs on Chinese wealth management firms, they found that contrary to displacement predictions, LLM adoption significantly increases advisor headcount. This effect is driven by a strong restructuring of human resources in wealth management firms, and LLMs actually stimulated hiring, particularly for junior and support roles.

They published their findings in International Review of Economics& Finance¹, under the title “Do large language models displace or complement investment advisors? Evidence from Chinese wealth management firms”.

From Routine Labor To Advanced Tasks

What makes LLMs a radically different type of automation is the type of task they can automate. Previously, automation mostly replaced routine manual labor, while LLMs can process unstructured information, perform logical reasoning, and generate complex texts.

So LLMs are more like complex, flexible robots on a production line than simpler, repetitive machines that can’t handle flexible tasks. And in the physical world, this kind of ability has led to a massive reduction in headcount in manufacturing. So it is not unreasonable for knowledge workers to fear being displaced by such technology: even if it does not fully replace humans, it could drastically reduce the manpower needed for a given task.

But automating away the task entirely is not necessarily the only effect of technology.

Another possibility is to make a previously complex task, only possible to handle by trained and experienced professionals, easier to do. In that context, LLMs could be more like early machines of the industrial revolution, turning a complex task requiring expert craftsmanship into a simpler manufacturing process requiring less training.

“By lowering the knowledge threshold required for independent task completion, such technologies may expand the range of tasks junior workers can handle autonomously.”

If that is the case, practical deployment of LLMs should have an effect not just on employment, but on the structure of the firms using them.

“Technologies reduce the cost of knowledge acquisition, production workers expand their problem-solving autonomy, escalate fewer routine problems upward, and increase the span of control of senior specialists, effectively flattening the hierarchy.”

Wealth Management, AI & Competitive Advantages

AI & Money Has A Trust Issue

On the day-to-day, employees at wealth management are performing a lot of routine tasks: market information gathering, report drafting, and compliance documentation. All of these are codifiable cognitive tasks where AI holds a clear comparative advantage.

However, the industry’s core value proposition rests on trust, fiduciary duty, and relationship management. These are interpersonal tasks requiring tacit knowledge and social intelligence that AI currently struggles to replicate.

This gap has limited AI deployment in the financial industry so far, as trust in AI advisors remains low.

AI Could Help Wealth Management Grow

At the same time, AI lowering the cost of financial advice could help the industry grow tremendously, as high advisory costs confine personalized services to high-net-worth individuals, leaving the vast mass-affluent segment chronically underserved when it comes to wealth management advice.

“If LLMs reduce the marginal cost of routine services sufficiently, they may enable firms to profitably penetrate this long-tail market, triggering AI-enabled scaling that drives service expansion and employment growth rather than displacement”

In that context, the Chinese market is especially interesting to analyze for the impact of LLMs on the industry:

• Rapidly expanding household wealth creates substantial latent demand in the mass-affluent segment.
• China’s CSRC and SAC enforce strict human-in-the-loop requirements, prohibiting algorithms from exercising full discretionary authority and mandating licensed advisor oversight of all personalized recommendations.
  • These rules channel AI adoption toward hybrid human-machine models, creating particularly favourable conditions for complementarity rather than complete substitution.
• China’s digitally-enabled financial ecosystem and widespread fintech infrastructure generate meaningful cross-firm variation in LLM adoption, enabling credible causal identification.

AI’s Impact On Wealth Management

What Can LLMs Do In Wealth Management?

The researchers distinguish three categories of tasks in wealth management that can be impacted by LLMs, each with a different level of substitutability.

The first set of tasks is technical analyses, including market data analysis, portfolio construction, investment research synthesis, and compliance documentation. Thanks to explicit rules, structured data inputs, and standardized outputs, these tasks are ideal for deploying LLMs to substitute human labor.

The second set of tasks is judgment and decision tasks, including risk assessment, scenario analysis, and personalized solution design. These tasks require integrating tacit knowledge and contextual awareness, so LLMs can only moderately substitute for human labor.

Lastly, there are the relationship management tasks. This encompasses trust building, emotional support during market volatility, and long-term relationship cultivation. These tasks require interpersonal sensitivity, contextual judgment, and authentic human connection, making LLMs mostly unsuitable in this role.

Gathering The Data

The researchers gathered multiple regulatory databases, corporate disclosures, and market data to construct a comprehensive panel dataset covering Chinese wealth management firms, adding to 7795 observations over 2019-2024.

To identify firm-level LLM adoption, job postings and employee profile information were collected from major recruitment platforms.

Wealth management personnel were divided into junior advisors (≤3 years of experience), mid advisors (4–10 years of experience), and senior advisors (>10 years of experience).

Finally, client-level analysis incorporates investment returns, asset allocation, fee structures, and client retention status.

Who Is Most Impacted

Previous studies showed that while LLM adoption in workplace settings revealed productivity gains of 14-56%, it disproportionately benefited less experienced and lower-skilled workers. As a result, firms placed greater emphasis on interpersonal and customer-facing abilities after adopting LLMs. However, demand for traditional technical skills did not significantly decline, suggesting that AI changed which complementary human capabilities firms valued rather than making technical expertise unnecessary.

“This pattern, where AI augments rather than simply replaces lower-skilled work, differs markedly from previous waves of IT adoption that typically complemented higher-skilled workers.”

The consequence is that LLM adoption is positively associated with increased emphasis on soft skills (communication ability, emotional intelligence, client relationship management).

Use Of LLMs

Only 17.8% of firm-year observations involve LLM adoption, indicating that while LLM use remains relatively uncommon, a meaningful proportion of wealth management firms have begun integrating these technologies.

Tech hub locations show larger effects than non-tech hub locations. This pattern likely reflects better access to AI talent, stronger technology infrastructure, and more sophisticated clients who value AI-enhanced services in major technology centers.

LLM use correlates most strongly with junior advisors, moderately with mid advisors, and most weakly with senior advisors, matching the expected profile of AI deployment in the wealth management industry.

When looking at what role LLMs are mostly used for, the study shows strong co-movement of LLM use with support roles and its near-zero association with client-facing positions.

Better Services, Same Prices

The study shows that LLM use is significantly positively correlated with all service quality metrics: portfolio return, risk-adjusted return, client retention, and AUM growth. LLM use is also associated with a decrease in complaint rate and regulatory action.

LLM use led to an 11.5% improvement in average portfolio returns. The 2.5 percentage point improvement in retention represents a 23.8% reduction in client attrition.

Still, there is no correlation between LLM use and advisory fees.

This likely indicates that competitive pressures may prevent firms from extracting higher fees, despite demonstrable service improvements, with benefits flowing primarily to clients rather than producers. This is particularly likely in an industry where service quality is observable, and firms compete on performance.

AI Does Not Replace But Restructure Human Labor

Overall, the findings show that AI restructures tasks in wealth management, but has no net displacement effect.

As LLMs assume routine analytical tasks, firms increasingly value employees who can exercise judgment (cognitive skills), work efficiently with AI tools (efficiency skills), communicate effectively with clients (social skills), and maintain strong client relationships (customer skills).

These findings do not suggest that technical skills are becoming unnecessary. Instead, they indicate that distinctly human capabilities become more valuable alongside technical expertise as AI assumes a greater share of routine analytical work.

By reducing the cost of technical tasks, AI could be not just a productivity tool but a powerful scaling tool for the wealth management industry, helping it expand its addressable market to a larger segment of the population.

These findings can extend to other industries with similar conditions: unanswered latent demand, human-in-the-loop regulation, and relationship-based service models.

Investing In AI Wealth Management

The study does not examine Morgan Stanley (MS ) or establish that the same results will occur outside China. However, it identifies the conditions under which AI-assisted wealth management may scale successfully. Morgan Stanley offers investors exposure to a business combining a large advisor network, self-directed investing, workplace services, and increasingly sophisticated AI capabilities.

Morgan Stanley

Morgan Stanley is a long-established giant of banking, lending, and wealth management, with $9.3T in total client assets in 2025, of which around 3/4^th are from the wealth management category. It is also worth noticing that total client assets have grown very fast since 2020.

Source: Morgan Stanley

Morgan Stanley’s primary robo-advisor platform is its E*TRADE brand, which charges an annual advisory fee of around 0.30%.

Beyond a pure robo-advisor, Morgan Stanley also provides other wealth management services:

• Virtual Advisors: Phone-based access to a team of remote advisors for structured planning without a dedicated full-service human broker.
• Full-Service Financial Advisors: Traditional 1-on-1 dedicated planning for complex wealth management needs.

Source: Morgan Stanley

This structure closely resembles the human-machine model examined in the study: AI handles information retrieval, meeting documentation, and routine preparation while licensed advisors retain responsibility for client relationships and recommendations.

As AI capabilities grow, it should allow Morgan Stanley to progressively reduce costs for technical skills and redistribute manpower toward a larger deployment of the human touch needed for more direct, personal advising.

Morgan Stanley’s size and reputation across all customer levels will help it deploy AI efficiently to millions of clients and expand the wealth management market.

It is worth noting that AI is already used as an entry point that channels customers toward higher levels of wealth management, with 7 million new client relationships since 2020, mostly funneled from E-TRADE and Workplace, a market-leading platform for financial advisors.

Source: Morgan Stanley

“Morgan Stanley at Work oversees approximately $1.2 trillion in assets through its workplace financial solutions business. The firm envisions corporate clients eventually engaging with the platforms “in a purely agentic way, using AI-powered tools rather than logging directly into ShareWorks or Equity Edge.”

So AI could prove a competitive advantage for Morgan Stanley through various mechanisms:

• Increasing the number of clients each advisor can support
• Extending personalized services further into the mass-affluent market
• Improving advisor productivity without reducing service quality
• Strengthening retention through more frequent and personalized interactions
• Scalingwealth-management revenue without proportionate administrative growth

This is not to say that AI will necessarily improve Morgan Stanley profits, as this study illustrates that such improvement and gains might mostly go to the customers. But it should help solidify its leadership position in the industry, and the likely global expansion of the wealth management market thanks to AI should benefit the industry as a whole.

Latest Morgan Stanley (MS) Stock News and Developments

Study Referenced

^{1. Lina Song and Huaili Lyu. Do large language models displace or complement investment advisors? Evidence from Chinese wealth management firms. International Review of Economics& Finance. October 2026. Article: 105636. Volume 111. 10.1016/j.iref.2026.105636}