Humans are ‘most important part’ says founder of AI-run asset manager
The Internet was becoming mainstream in the late 90s, but Miro Mitev was trying to explore something that would not become popular for decades: Artificial Intelligence.
Mitev, now an asset manager, became an early adopter of artificial intelligence in finance after discovering the capabilities of neural networks while studying at the Vienna University of Economics and Business in 1997.
He told CNBC that he sees the potential of neural networks for financial forecasting. “I fell in love with those kinds of possibilities,” he said.
Mitev has spent his 25-year career making predictions for banks and technology firms like Siemens. He founded SmartWealth Asset Management, whose decisions are made entirely by a network of artificial intelligence systems. Its latest fund, IVAC, targets $2 billion in assets under management and has a target annual return of 14-15%.
Although there is no human intervention in the AI’s decisions, Mitev said “humans are the most important part of the equation” as they are the ones selecting the training data, inputting variables, creating parameters and constantly fine-tuning the model.
Once a model is established, “it is very dangerous to start intervening,” Mitev said. In fact, he added, trusting the model is his golden rule.
Instead, people must ensure that there are no errors in the data or calculations and introduce new data to keep the model up to date.
“The worst thing is to invalidate the results, and this is something that happens very often,” Mitev said, adding that people “don’t trust” AI at first. “As humans, even if we can’t see the outcome right now, when we look back in two months, three months, we say, ‘Actually, we were wrong,'” he added.
The forces that drive the market (optimism, pessimism, and speculation) are deeply human. Even the European Central Bank has warned that the current AI bull run may be due not to detailed technical analysis but to fear of missing out.
Mitev said removing emotion when investing yields better results; SmartWealth Asset Management has gained 407.63% in the 10-year period through Nov. 1, 2025, according to a chart a representative of the firm shared with CNBC; This rate was an industry indicator of 145.34% in the same period.
Mitev said it is “not possible” to know what will happen in a year, but he can see up to a month ahead with his model. “Evaluating this information and making informed decisions based on it is proving to consistently deliver better-than-human outcomes.”
Considering that artificial intelligence systems “hallucinate”, that is, produce false information, continuous monitoring and the introduction of new data are important points. Errors in the models were caused by “overfitting,” data problems or misspecification of the model, Mitev said.
Overfitting is when the algorithm pays too much attention to what Mitev calls “noise.” He said this was “non-significant” data because it did not reveal a true cause-and-effect relationship with stock performance.
Mitev added that rigorous design, validation, and in vivo testing serve as an antidote to this. This means that although the fund strategy is driven entirely by a set of algorithms, humans still play an important role in ensuring it is effective.
“This is actually a process that has evolved over the years… and that’s why it’s so important to develop these types of technologies in-house,” he added; especially for anyone who wants to make a difference in artificial intelligence games.
