Technology
Towards Optimal AI-based Wealth Management - Study

The AI impact
As shown in the research conducted by Accenture [6], there is
approximately $78 trillion of assets ready to be captured by
wealth managers (due to the global expansion of the middle class
and wealth created by a new generation of entrepreneurs, e.g.,
those who decided to embark into their own business thanks to the
great amount of information available on the web and the almost
zero cost of reaching customers through social media). For this
reason, AI presents a good fit for targeting this market since it
provides:
-- Major client engagement through the use of web-based
platforms by the advisors and their own clients;
-- It helps to elaborate better financial products such as
portfolio optimisation by using machine learning;
-- User experience is enhanced by providing a transparent
24/7 readily available source of information in a website;
and
-- AI increases productivity and operational efficiency
since the big majority of the tasks are performed by AI-automated
systems (e.g. portfolio allocation given the clients preferences,
automated order placement and free access to all accounting
services).
The adoption of AI is not reserved for fintech start-ups. There is a clear adoption by major institutions in the market proving its fundamental value to address the new market challenges. Actually, there is a trend of big corporations incorporating AI-based solutions in their investment and portfolio allocation repertoire. Examples include Abrdn, which recently acquired Exo Investing [7], a move that is intended to deliver a 24/7 digital wealth management solution via an app and JP Morgan has also bought another fintech firm: Nutmeg (containing approximately £3.5 billion in assets under management for more than 140,000 clients) [8]. Perhaps, one of the most successful stories of AI in wealth management is the case of The Next Best Action system by Morgan Stanley, which provides their financial advisors with machine learning algorithms to identify investments of interest to particular pre-existing clients [9]. From this practice it has been shown that continual engagement with the client has improved the overall experience and motivated substantial valuable changes in their winning strategy.
Machine learning made simple
Machine learning can be seen as a subfield of AI concerned with
the incremental learning of artificial systems from data with the
central objective of taking advantage from previous experience.
AI/ML makes the investment process better by systematically
making an abstraction of the wealth management process and
transforming it to a pipeline of the following automated
tasks:
-- Preference profiling. Smart front-end interfaces gain insight
into the current client situation by providing an automated
questionnaire which keeps track of the answer history and then
mathematically transfers this information into a classification
process for user profiling. For example, in terms of risk
tolerance we can segment clients into cautious, balanced or
aggressive. Using transfer learning, we can also significantly
reduce the amount of time it takes to complete these
questionnaires as one of the primary characteristics of younger
generations is lower tolerance for completing forms;
-- Asset allocation. Based on previously trained models and the
client's profile, an AI-based system infers an optimal solution
for the allocation of wealth by using a predefined portfolio or
dynamically tailoring a new option for covering specific needs.
From our previous example, cautious clients are immediately
assigned a portfolio with a large majority of fixed income
securities, a small proportion in equities and a minimal
proportion of cash and equivalents. Balanced clients are
automatically assigned portfolios with an equal amount of fixed
income securities and equities. Aggressive clients take a minimal
proportion of fixed income securities and a major part of
equities, keeping a minimal amount of cash and equivalents;
and
-- Order management. Clients can opt for a fully-automated
solution that places orders in the market autonomously, or they
can impose stricter controls for order approval.
What’s next for wealth management?
My vision of the wealth management sector of the future involves
the construction and development of data-driven machine learning
solutions. Specifically, extending the notion of modern portfolio
theory by driving the investment process through the use of
automated AI-based systems for asset allocation, order management
and placement, reporting and portfolio analysis. Clients of the
future are -extremely- tech savvy, therefore they should be able
to enter a holistic application designed to meet their needs, and
at the same time being accessible from any computer, mobile
device or tablet.
Disruptive technologies should aim to revolutionise the
investment process in wealth management, providing an automated
combined solution offering:
-- High returns over a low cost. The new business model
should use a data-centric paradigm where machine learning
algorithms are totally in charge of automated asset allocation,
supplying conventional human intervention in portfolio creation
(having a proven performance over passive ETFs offering
uncorrelated portfolios to major indices reducing risk).
Web-based fund monitoring and accounting tools make clients
totally independent in any reporting or order management tasks.
-- Full transparency. Automated solutions should provide
the client with full 24/7 access to the most detailed information
regarding allocation, exposure data and portfolio risk.
-- Excellent client experience. Clients should be allowed
to gain instant access to their data taking advantage of high
levels of automation, efficiency and mobility on demand.
-- Tech-driven advice (fully or partially automated). Full
automation produces an optimal tailored portfolio given a
personalised requirements elicitation process. Furthermore,
direct communication to the client enhances the investing process
by aligning those automated recommendations to special requests
by the clients (e.g. interest in a sustainable ESG approach,
risk-aversion level modification, or a different rate of return).
Integration is paramount. Currently, incredible efforts need to be put in place in order to integrate several service providers and their outputs to access a portfolio management system to keep track of performance and exposures; a risk management system to visualise historical risk-metrics (volatility, Sharpe ratio, etc.) by considering benchmark indices and performing factor analysis in order to statistically explain the nature of the returns; an order management system to review and control any order to be executed as well as keeping a history of previous orders for reporting purposes; an information management system for having direct access to all the relevant information about their investments and a data lab to allow them to experiment with back-testing scenarios of their strategies.
The use of AI in investment management is set to revolutionise the industry. A disruptive holistic approach described in this paper fills the gap between end clients and targeted performance from their portfolios by automating the entire investment process. Financial advisors need to augment their skills with the advent of the new trend of technologies in order to have a competitive advantage [10]. Operational costs can be highly reduced by opting for a fully-automated solution.
The future is bright. I am optimistic that for these new generations of investors a well-deserved and trustworthy set of opportunities will (and can only) be offered through innovative technology.
About the author:
Dr Sonia Schulenburg is director, and investment committee member of Level E Capital SICAV plc, a Maltese multi-fund investment company dedicated to autonomous investing. She holds a PhD in Artificial Intelligence from the University of Edinburgh and a BEng in Computer Engineering (summa cum laude; 1st Class Honours) from the ITAM in Mexico City, a Professional Certificate in accounting from the University of California, San Diego and a postgraduate degree in Corporate Strategy and Finance from Edinburgh Napier University, where she graduated with distinction in both.
Acknowledgements
We would like to thank Steve Dyson from Investment & Wealth
Management Consultants Ltd for the interesting conversations,
support and guidance while conducting this research paper.
References
[1] Financial Advisor Fees Comparison – All-In Costs For the Typical Financial Advisor? Kitces.com Website. July 31, 2017. Accessed on August 25, 2021. https://www.kitces.com/blog/independent-financial-advisor-fees-comparison-typical-aum-wealth-management-fee/.
[2] Investors Want More Diversified Financial Products and Customized Advice from Their Wealth Managers, Accenture Report Finds. Business Wire, A Berkshire Hathaway Company. Online Article. August 24, 2021. Visited on August 25, 2021.
https://www.businesswire.com/news/home/20210824005209/en/Investors-Want-More-Diversified-Financial-Products-and-Customized-Advice-from-Their-Wealth-Managers-Accenture-Report-Finds.
[3] The Future of Wealth Management. The Street, Retirement Daily. Online Article. August 23, 2021. Visited on August 25, 2021. https://www.thestreet.com/retirement-daily/financial-adviser-center/the-future-of-wealth-management.
[4] Thematic Research into Wealth Management – Players Include DBS, Betterment and UBS Among Others –. Research and Markets, The World’s Largest Market Research Store – Yahoo Finance. Online Article. August 17, 2021. Visited on August 25, 2021. https://finance.yahoo.com/news/2021-thematic-research-wealth-management-093800094.html.
[5] 10 Disruptive trends in wealth management. Deloitte technical paper. Accessed on August 25, 2021. https://www2.deloitte.com/content/dam/Deloitte/us/Documents/strategy/us-cons-disruptors-in-wealth-mgmt-final.pdf.
[6] AI in wealth management: Built to scale. Accenture Capital Markets. Online Article. December 02, 2020. Visited on August 25, 2021. https://www.accenture.com/gb-en/insights/capital-markets/wealth-management-artificial-intelligence.
[7] Abrdn acquires AI solutions business Exo Investing. Investment Week. Online article. August 10, 2021. Visited on August 25, 2021. https://www.investmentweek.co.uk/news/4035619/abrdn-acquires-ai-solutions-business-exo-investing.
[8] JP Morgan buys Nutmeg. Fund Europe. Online article. Accessed on June 23, 2021. https://www.funds-europe.com/news/jp-morgan-buys-nutmeg.
[9] The Pursuit of AI-Driven Wealth Management. MIT Sloan Management Review. Online Article. July 07, 2021. Visited on August 25, 2021. https://sloanreview.mit.edu/article/the-pursuit-of-ai-driven-wealth-management/.
[10] AI won’t replace investment managers but it could improve
returns. World Economic Forum. Online Article. May 24, 2021.
Visited on August 25, 2021.
https://www.weforum.org/agenda/2021/05/ai-wont-displace-investment-managers-but-it-could-improve-returns/