- July 1, 2023: Vol. 10, Number 7

AI and the investing process

by Benjamin Cole

Download Audio file

By mid-2023, no acronym on Wall Street had become hotter than “AI” for artificial intelligence. Not only were investment managers scouring the world for AI plays, but financial advisers were increasingly touting AI as a means for boosting investor results.

The expression “the future is now” may be a cliche, but already serious financial advisers, including the Big Four accounting and advisory giants, are touting AI services to investment houses to refine their money-placing strategies.

Investors, small and large, have long used “screens” as a first run to cull promising stocks, such as selecting all stocks with price/earnings ratios below 10, three years of profit growth, and market caps above $1 billion. From there, hands-on and laborious analysis might take place, the reading of dozens of filings and all news articles that might have made it into financial media.

AI gurus say there is a better way. “Analyzing data from 10 years of financial statements for dozens of companies in a typical fund manager’s portfolio is the stuff of nightmares for research analysts — yet it is the perfect learning opportunity for AI,” according to  Thierry Grouès, Paris-based associate partner with Ernst & Young.

Certain patterns might emerge through such an intense AI analysis, as well as buying or selling trigger points that hitherto were not recognized, reported E&Y.

In truth, many large investment houses or brokerages end up with too many stocks to follow, if due diligence is to be constantly conducted on each one. The “span of control” for any particular human analyst is limited.

At the same time, it is hardly a secret that Wall Street is becoming increasingly digitized and online. That means AI, and another acronym, RPA (robotic process automation) can help portfolio managers continuously review stock performance, public filings and disclosures of portfolio assets, and then “ring a bell,” or even automatically buy or sell, when certain conditions are triggered. Emerging patterns, unseen to humans, can also be flagged.

An undisclosed client of E&Y, with billions under management, recently “needed to accelerate and improve the decision-making process by using AI to interpret larger and more complex data sets and, in some instances, automate decision making altogether,” according to Grouès.

Operating at behemoth scales, institutional Wall Street is quick to grasp at even the smallest edge in investing. After all, when assets under management (AUM) often soar into the hundreds of billions of dollars, even a 0.2 percent boost in investment outcomes can result in additional millions in returns, which are then compounded into the future. Not only that, some investment managers have literally trillions under management, such as BlackRock, which in 2022 passed more than $10 trillion in AUM. The math is compelling: 0.2 percent of $10 trillion is $20 billion. That’s in one year. Obviously, a few extra billion dollars of return annually can justify a large amount of AI.

Not surprisingly, the hottest academic degree on global Wall Street right now is not an MBA but one in data science.

“Leading companies are hiring hundreds of data scientists and using cloud infrastructure to further increase the power of AI,” says Grouès.

Institutional Wall Street has discovered and is implementing AI. But, what really defines AI in the investment world?


There is no single definition of AI deferred to by all, but Deloitte, the Big Four accounting and consulting firm, posits this description: Artificial intelligence is a suite of technologies, enabled by adaptive predictive power and exhibiting some degree of autonomous learning, that dramatically advance our ability to:

  • Recognize patterns
  • Anticipate future events
  • Create good rules
  • Make good decisions
  • Communicate

AI may be particularly useful in exploiting, in a timely manner, information that may be beyond the normal daily attention-zone of many analysts. For example, AI can tirelessly examine alternative datasets, such as weather forecasts, smartphone GPS patterns, containership movements, or search engine word frequency, to help structure investments, advises Deloitte.

Recently, an investment analyst was provided AI-generated information of smartphone-GPS patterns around various U.S. downtowns, showing foot traffic trends in the post-pandemic era. (Cellphone companies can sell “sanitized” data of smartphone GPS-locations, serving as an excellent proxy for foot traffic.) AI can also take very deep dives, based on years or even decades of data, into “relationship mapping” or identifying nonintuitive relationships between specific securities and economic or broad-market indicators.

Yet as sexy as AI and investing sounds or may become, the real value of AI to investment managers may be in back-office operations, according to Deloitte. Among the many tasks that will become increasingly handled by AI-enhanced algorithms will be the generating of reports for clients, including portfolio and risk commentary, and related marketing material “using natural language processing,” says Deloitte.

A little unsettling for some perhaps, Deloitte envisions investment managers using chatbots and machine-learning to respond to employee or investor queries, while concurrently generating related reports for management. The “monitoring of suspicious transactions,” can be automatically flagged.

And AI will also offer to management “employee insights” including the monitoring of “employee conduct risk and employee morale.” One might well wonder if AI will be tuned into investment manager company emails and texts, or on-location company webcams.

A brave new world to be sure, perhaps with a hint of the Orwellian, but AI ultimately will likely lower costs of operation for money managers, resulting in lower and more competitive fee structures. Investment managers burdened with too many human mouths to feed may find themselves priced out of business.


Just as the internet brought a Niagara of information to any at-home investor with a wifi connection, so now AI and the web may arm every Wall Street watcher with AI, even if not an institution or a brokerage client.

In May, readers of financial pages were treated to headlines such as this from Fox Business: New AI-powered investing tool uses ChatGPT to manage portfolios.

Touting a Chatbot GPT app “plug in,” an enterprise named Portfolio Pilot says it brings AI management to even retail investor stock portfolios.

For the uninitiated, ChatGPT is an online AI chatbot that uses natural language processing to engage in human-like conversations. But the AI-power behind Chatbot GPT can also be directed to manage portfolios, according to Portfolio Pilot.

Portfolio Pilot’s AI portfolio assesment “analyzes thousands of factors” and they tout a “3 percent to 6 percent boost in returns.”

Portfolio Pilot has plenty of competitors: In late May, The Armchair Trader, a U.K. portal for retail investors, disclosed a partnership with Israeli AI-developer Bridgewise to provide AI-generated stock analysis to community members.

The AI-for-retail investors offerings raise a challenge for the entire financial advisory industry.

The future is now, and retail investors can trade through very low-cost online brokerages and have portfolios reviewed constantly through AI-assisted algorithms. As the retail investor provides more information (age, marital status, anticipated income, geographic  location) the AI-adviser can offer tax and retirement planning wisdom.

With AI emerging but likely set to grow, the question is inevitable: What can human advisers offer retail investors that will validate the costs?


Most likely, except for AI proponents, the majority of Wall Streeters would say the jury is still out on AI, or if AI does work, then one must wonder how long until much of the market is “AI-ified” and advantages are minimized.

However, in March, Hyungjin Ko and Jaewook Lee, both of Department of Industrial Engineering, Seoul National University, produced a study entitled Can ChatGPT Improve Investment Decision? From a Portfolio Management Perspective. The pair said they compared the performance of portfolios constructed by ChatGPT to portfolios built on randomly selected assets. The ChatGPT’s portfolios won.

Ever since the classic finance book by Burton Malkiel, A Random Walk Down Wall Street, was published in1973, the idea of the efficient market hypothesis has largely prevailed, in theory and in practice on Wall Street. In a nutshell, very few, if any, money managers consistently beat the market or equity investments selected by throwing a bunch of darts at the stock tables of The Wall Street Journal.

Yet, Ko and Lee said their analysis “indicates that portfolios constructed using ChatGPT’s selections outperform those constructed using randomly selected assets.” The ChatGPT model was able to generate greater diversity through intelligent selection, than the random-darts approach. The more robust diversity, as might be expected, lowered relative risk in the ChatGPT portfolio, said the authors.

“Regarding minimum risk portfolios, ChatGPT-based portfolios exhibit lower risks than those of randomly selected assets, demonstrating that ChatGPT can significantly enhance the overall risk management of the portfolio,” reported Ko and Lee.

One study from academia is hardly the last word on AI-driven investing. Nevertheless, there is some sense to the idea that AI-run portfolios can more tirelessly, rigorously and consistently adjust portfolios to obtain greater diversity than human-run portfolios or random chance.

For large investors of scale, every small advantage will be seized. If AI delivers lower risk for similar returns, then AI will be embraced.


As more data and information is digitized and online, the more compelling the arguments for AI become, in all fields, including investing.

An interesting era is likely about to be entered into, and that is the age of AI-driven investment management. However, Wall Street history tells us technical advantages gained are fleeting and are quickly equaled by market adaptation.

It may be that positive AI-driven investment results fade with time, as more investment houses turn to intelligent robotic running of portfolios. As many a pundit knows, no one beats the market for long, and for decades the giant index funds have more than held their own against managed funds.

Yet, in another scenario, perhaps Wall Street winners will emerge with large investment houses that can afford increasingly powerful AI-driven investment strategies, while those who cannot and are left behind. Already, investment managers are colossal, measured not only in billions now, but even trillions of dollars under management. Further industry consolidation may result from investment managers using AI not only to win as investors, but keep employees and clients better apprised of operations — and if AI lowers costs of operations, then the more expensive non-AI operated investment managers will be under constant pressure to cut costs while delivering better results.

Perhaps there will be another group of winners, investors who return to deep analysis of a few companies, learning management teams in an industry in which they are already deeply knowledgeable, perhaps through years of employment.

One can at least imagine there will be niche markets in which undigitizable information, the kind not found online, provides a leg up for the patient and shrewd investor.

But as the giant investment houses are capturing larger market share and are likely to adopt AI if only to lower operational costs, the future on Wall Street looks intelligently artificial.


Benjamin Cole ( is a freelance writer based in Thailand.

For reprint and licensing requests for this article, Click Here.

Forgot your username or password?