Originally published in Italian on Milano Finanza
written by Marco Capponi
English translation:
MDOTM uses Artificial Intelligence to assist institutional investors in asset allocation and security selection, working throughout the entire process. However, it never aims to replace the role of human beings. Try asking ChatGpt to construct an investment portfolio that adheres to specific parameters, such as risk profile or the subscriber's preferred asset allocation. The results are, at the very least, questionable: in terms of financial knowledge (a geographic fund on Piazza Affari includes Atlantia and Nestlé), and even regarding the most basic risk concepts. It can happen, for example (this newspaper has experienced it), that a small investor with a medium risk profile is advised – albeit with the necessary note of consulting an expert before making any decision – to allocate 20% of their wealth to cryptocurrencies. To understand the benchmark, a recent Credit Suisse report estimated that only 2% of cryptocurrencies in an ultra-wealthy individual's portfolio represents a quarter of the total risk.
It's not gambling.
In short, trusting the current ChatGPT to invest is akin to putting your money into a slot machine. Pure chance. However, it is important to understand a fundamental point: Artificial Intelligence is not just ChatGPT "OpenAI's software specializes in generating text, while our AI specializes in understanding and analyzing markets." Speaking is Federico Invernizzi, Chief Operating Officer of MDOTM, a London-based scale-up (with an office in New York and one in Milan) that had the insight long before ChatGPT became popular: to create an AI-based platform (Sphere) to support institutional investors in building their portfolios. Among those who have chosen to collaborate with MDOTM are the Sella Group, which recently launched an AI-based wealth management line, as well as UniCredit, Lazard, and Momentum Global.
Blank slate?
No, thank you. Despite the easy allure, asset managers can rest easy for a long time: Artificial Intelligence alone, starting from a blank slate, is not capable of constructing an investment portfolio. "Or rather," clarifies Invernizzi, "theoretically it could, but without a chief investment officer who chooses the methodology and objectives of the fund, while also mediating with clients, the work of AI would make little sense to exist." Instead, Artificial Intelligence left completely free can have a supporting role: "It is capable of providing a completely neutral and unbiased view of the world in which one invests: however, it is up to the human being to decide whether it makes sense to follow that view or not."

AI-Driven Portfolios
What MDOTM is already doing is "bringing some AI into the investment process, without replacing the human role." The great advantage of machines in this context is their ability to process immense amounts of data in a very short time, performing tasks that would be impossible for a human manager due to obvious limitations of time and scale. "Essentially, AI can be used in three ways: upstream of the investment process, meaning the ability to generate excess returns through correct positioning in the investment world; downstream of the process, i.e., the actual construction of the portfolio given a certain market scenario; or in a hybrid form between the two models." It is, the manager adds, "a series of tasks that managers have always performed and that constitute an essential part of any portfolio construction: AI is complementary and simplifies the work."
The upstream scenario
The starting point is the definition of the market scenario, the alpha intuition. "In this context," highlights Invernizzi, "the inputs fed into the software are the preferences in terms of expected portfolio risk, the reference time horizon, the asset classes in which one wants to invest, the active and benchmark portions of the portfolio." What can be expected at this point? "AI can outline both the asset allocation and the over or underweighting of certain asset classes." But it can also go deeper, "for example, by hypothesizing expected returns."
The downstream process
If, on the other hand, there is already a defined market scenario and one wants to determine what to actually include in the portfolio, operational alpha comes into play. "In this case, the input is the set of expectations from the scenario: for example, a geography that will perform particularly well or poorly, or an over or underperformance on government or corporate bonds." In this case, the machine "outlines the optimal portfolio for the set scenario. So, if one assumes that a geographic area will perform well, what are the securities to include in order to benefit from this performance?" For every worldview, essentially, "there is a way to implement it, which can extend to individual securities, funds, or ETFs to include in the strategy. This process can be carried out by AI," concludes Invernizzi.
(reproduction reserved)

--------
About MDOTM Ltd
MDOTM Ltd is a global provider of AI-driven investment solutions for Institutional Investors. Founded in London, the company has a team of over 60 physicists, data scientists, engineers, and finance professionals. MDOTM Ltd provides Portfolio Advisory and Asset Allocation services to Banks, Insurance Companies, Family Offices, Pension Funds, Wealth and Asset Managers. Numerous financial institutions across the UK, Europe, and the US leverage MDOTM Ltd’s advisory and AI platform – Sphere – to support their investment process. For more information, visit www.mdotm.ai and www.mdotm.ai/openai
Contacts