WealthBriefing talks to the chief executive and founder of a Swiss asset management firm that is using AI to guide stock and bond investments.

When markets jump around – for example the crash in early March 2020 amidst the pandemic or rebounding after the April 2025 selloff – investors who succeed in avoiding the worst losses and who can capture returns have a reason to be pleased.

In all the ways that portfolio managers try to avoid the “noise” of media coverage of current events and other forces, one is the approach that taps into the science of neural networks, which is an important element of artificial intelligence. These days, it is argued, you need AI to help avoiding mistakes. Above all, taking human emotion out of the mix is important.

Dr Miro Mitev (pictured), founder and CEO, Smart Wealth based in Zurich, is using AI to automate investment management. In July, the firm launched its Irish Collective Asset-management Vehicle (ICAV), an umbrella structure with two sub funds: the SW Multi Asset AI Flagship Fund and the SW Global Equity Plus AI Fund. The ICAV expects more than $100 million from investors from Europe, the Middle East, Asia, and Australia. The structure aims to secure $2 billion in assets over the next three years. Seed share class investors include family offices around the world.

As explained by Smart Wealth on its website, “every purchase and sale in our AI-controlled investment process is based on data and facts,” completely stripping out the human element and risks of emotions entering the mix. The firm’s forecasting models handle variables such as macroeconomic, fundamental and technical inputs of different frequencies. When tailoring a strategy to match a client's requirements, a client can, for example, choose what industries they want to include. The idea is to build an “optimal portfolio” with the highest return for a given risk level.

Dr Mitev is ambitious about what Smart Wealth can achieve.

“We want to be a core asset manager for the industry worldwide. We want funds to be in the multi-billions [AuM] in three to five years,” he told WealthBriefing in a recent interview. 

“Our system is not built on news…our system is not reacting to noise but to hard data and causal leading indicators…it looks at what information is significantly important and what is not,” Dr Mitev said. “It identifies turning points and the probability that a future change of direction is very high.”

“This is very robust in different market cycles,” he said. 

Dr Mitev said that to demonstrate how it worked, just before the global pandemic broke, Smart Wealth shut 60 per cent of its portfolios two weeks before markets collapsed in early March 2020; it later restored portfolio positions one and a half months later. On the Russia/Ukraine story, it moved into cash on 3 January 2022 by reducing its bonds exposure and on 25 January 2022 cut its equity`s exposure to avoid some of the most significant market setbacks.

Using this process, for example, a portfolio might move into cash for a particular period.

Smart Wealth Asset Management – to give its full name – has $400 million of AuM, and a team of 30 professionals across Europe, the Middle East and Asia. The firm said it is a consistent top tier performer; its model portfolio has beaten its composite index by 6.5 per cent per annum.

Smart Wealth gives this chart showing top-tier performance, until July this year, among prominent multi-asset products. The white line has Smart Wealth at the top.

(Data source: Smart Wealth, Bloomberg.)

The AI impact
The rise of investment houses such as Smart Wealth underscores the growing use cases of AI, although immense computer power has been applied to trading systems for years. 

Dr Mitev has more than 25 years of experience in what can be called the quantitative area; he started his involvement with AI in the late 1990s. He has worked in investment banking and asset management and built strong experience in applied research in areas such as operations research, statistics and portfolio optimisation. Dr Mitev wrote his master’s thesis on neural networks and their applicability to the stock market, applying his insights to stock market investing. 

And what did the evidence show? Via empirical data, Dr Mitev said he found that using a neutral network approach improved the accuracy and reliability of stocks forecasting. 

Over the years, Dr Mitev developed trading technology and insights for banks and asset managers. There is also an academic side: he has taught at Vienna University of Economics and Business Administration.

Making new data-driven approaches can work today because technology makes them possible. 

It is easier to develop such quant models today because data and computational speed is more available and the technology that can handle it is more sophisticated and powerful.

Stocks and bonds
Smart Wealth only works in the areas of blue-chip listed equities and stock/bonds ETFs. The firm provides investment products to institutions and professional investors, including family offices.

Firms that have used Smart Wealth's forecasts include JP Morgan, Commerzbank, Fidelity Investment Services, Merrill Lynch, Hyundai Asset Management, Nomura, Société Générale, and Santander Asset Management. 

Dr Mitev says interest has been positive: “They [users] want double-digit returns and they want daily liquidity…an annualised return of 14 to 15 per cent.” Private banks and independent asset managers are interested in what SW is doing; there has been “strong” interest in SW funds, he added.