By Nayan Ramani, product manager at Google

It’s not just consumer retail websites or manufacturing processes that employ Artificial Intelligence (AI) and Machine Learning (ML) to sift through data to look for trends and instructions on the next steps to be transacted. Increasingly, the world of investment and portfolio management, the companies that help grow wealth for their clients, are taking advantage of AI. Whether to attract younger investors who are more comfortable with making financial decisions via an app on their phone, to help generate diverse investment ideas for investors, or to keep human employee costs down, the use of AI results in investment and portfolio management that is less expensive for firms and more accessible for customers.

While the use of AI and ML in the investment and portfolio management sector is expanding, there is plenty of room for growth in the future from both the customer end and the investment manager’s side of the equation. As per the 2019 Deloitte report,  use of AI applications enables the investments firms to generate better alpha (returns on investments), improve operational efficiency, customize product delivery and manage risk.

AI has “democratized” the access to professional wealth management for the end users, making it accessible to more people, especially those in the younger demographics who are more comfortable with managing their investments via robo-advisors or online platforms like wealthfront.com. This option has greater appeal than taking the time out of their day to sit across from an advisor face-to-face. For this target market, it’s easier to enter some parameters, answer a few questions (age, income, risk tolerance, retirement age goal, etc.), and create a portfolio that is customized to their investment preference and risk profile.

For investment managers and financial advisors, AI can help achieve a higher level of performance by enabling smarter and faster decision making for a much larger population of clients, thereby allowing them to reduce service fees per client to make the service more accessible. Moreover, the AI and ML routines let advisors scale up their own levels of production. Even when the workday is over, a customized and automated AI advisor can continue to monitor the news events related to customer portfolios, assess the impact, positive or negative, and raise an alert for the human advisor to take an action or take an action on its own.

How financial institutions can reap the benefits of AI

One of the mostly costly aspects for the traditional wealth management firm is building and staffing a branch with employees who are there to answer customer questions. What if those same questions could be answered via an app, at any time of the day or on weekends by a “digital assistant”? This will lower overhead costs, freeing up the advisor’s time, which then can be used for more value-add activities like investment analysis or improving the AI algorithms themselves.

AI can also make it easier to establish a long-term relationship with customers and create up-sale opportunities: an algorithm that sees a 22-year-old just out of college approved for a car loan may prompt messages about favorable rates for a mortgage a few years down the road. Creating these personalized opportunities for clients (and cross selling different financial products) are easier with AI. Advanced digital assistants that are using “Natural Language Processing,” to sound more like real people in their responses to queries also make the app user feel comfortable about choosing not to sit across a table from a human when making decisions on their financial future.

Important considerations when using AI

As with any application that relies on the data recorded to make decisions via an algorithm created to sift through that information, it’s all about the quality and integrity of the “training data.” Poor quality of training data often leads to “garbage in, garbage out.” There can also be bias if the training data is not diverse enough to cover a variety of scenarios – skewed results and poor decisions being the outcome. As recommended in a 2019 BlackRock report AI and ML technologies with access to sensitive data must also feature “robust” cybersecurity defenses that include data encryption, cybersecurity insurance, and incident management frameworks. As for risk management, the loss of trust and a client’s money impacts the institution’s bottom line as well as reputation, creating unwanted headlines and giving competitors a leg up. Bad data and a faulty algorithm can lead to sub-par returns on investments which, in turn, results in a loss of commissions.

Here’s where AI can help. When it comes to risk management, AI assists with both market risk and credit risk. Market risk is the possibility that a portfolio will experience losses due to factors that affects entire financial markets, like recession, geo-political issues, natural disasters etc. AI algorithms can scan through textual or image data sources like news articles, central bank press releases, financial contracts etc. and help incorporate such qualitative data into risk modeling to produce the accurate forecast of critical market risk variables like Value at Risk and bankruptcy probability. Credit risk is specific to a company and results from failure to repay a loan or meet a contractual obligation. Portfolio managers need to monitor the credit risk of individual holdings as well as the entire portfolio. The larger the number of positions in the portfolio, the more complex it is to monitor the credit risk. A combination of AI techniques like artificial neural networks (ANN) and support vector machines (SVM) can be used to forecast the credit ratings and the probability of bankruptcy. Also, ANN and SVM can be used to estimate the loss given default (LGD), an important measure of economic loss when default occurs. An algorithm can also pore over years of financial statements to help inform investment decisions, such as is the company on a constant roller coaster or has there been steady growth over time? Customized algorithms can produce the reports needed on a regular basis to stay fully apprised of changes in market conditions and provide immediate alerts to remain in the loop on those changes.

A lack of transparency on how AI-based algorithms make financial decisions for a customer is another consideration and raises important questions. These include:

• Who should be blamed for a poor investment decision based on a faulty recommendation that came from an algorithm?
• Where do the fiduciary and legal responsibilities rest?
• Can investment managers say it is out of their hands if AI was employed and the customer was aware of that methodology?
• What if there is no physical presence for an online-only financial firm?
• Who is ultimately responsible if a third-party provider is hired to set up the AI operations for an investment management company?
• What if faulty data is shown to be at the root of a portfolio loss?

These concerns may not be major issues yet in the investment and portfolio management world, but as AI becomes widely used in the financial sector they need to be addressed. The line of responsibility is still unclear at this point, even with the waivers clients agree to when they onboard.

AI’s place in the investment management world     

There are many in the investment and portfolio management world still coming to grips with how to employ AI and ML with their clients. It’s a matter of building a level of trust with clients and helping them understand that the analyzed data and automated decisions are based on good data and expertly crafted algorithms designed to make sense of that information. The Gen Z and millennial age groups that are more comfortable with AI agents and mobile apps will help hasten the further penetration of AI and ML into the portfolio management sector. In 2019, The CFA Institute Foundation reported that only about ten percent of asset managers were employing AI and ML, with more anticipated to use that technology in the future. When the scope is widened to include banks, insurance, and other financial services, nearly sixty percent are on board with using AI and ML at some level. Financial institutions that do not adapt facets of AI or develop algorithms that can mine data to create customized product offerings will be left behind and may never to catch up with competitors constantly updating their own digital footprint.

Incorporating AI as part of the cost of doing business also means hiring talented programmers that soon, if not currently, will be in great demand and mostly likely will come with a high price tag. This is a service many customers, especially in younger age group demographics, already expect in other experiences, whether ordering meals via an app or going online to buy products for the home. This begs the question: why are they not able to seek approval for a customized loan or invest in a company’s stock the same way? Smaller investment management firms without the capital to support AI and ML technology and the talent to make it all work could find themselves falling by the wayside or having to join forces with larger firms to continue serving clients they have been helping for many years. As in many other business and consumer sectors, these digital transformations will continue to alter and shape the playing field for investment and portfolio managers in the coming years.

About the Author:

Nayan Ramani is a product manager at Google with expertise in capital markets, prime brokerages, financial risk management, and payments infrastructure. He has over 14 years of experience with fintech systems. He can be reached at [email protected].