Local fintech company NMRQL Research has launched what it describes as South Africa's first machine learning-powered unit trust, using artificial intelligence to better drive stock selection.

Unit trusts are asset portfolios such as equities and bonds, that investors and investment firms can put money into.

NMRQL Research, however, will put a twist on this by investing in a diversified portfolio of domestic and international assets, with asset allocation and stock selection systematically managed using machine-learning algorithms.

The machine learning-powered, computational investment process is designed to allow the company to discover hidden patterns in underlying big data, which can then be exploited to forecast returns across all asset classes and markets in a bid to secure long-term growth in capital and income.

The company is co-founded by former FNB CEO Michael Jordaan, who said the launch of the fund marked the start of a "paradigm shift" in the South African investment management industry.

"In addition to the vast amount of data that the algorithm is able to process, the investment philosophy eliminates emotive decision making, which allows the model to remain rational at all times," he said, adding that humans have "cognitive biases" that negatively impact on objective reasoning.

"These are compounded when financial repercussions are involved," he said.

NMRQL is headed up by CEO Tom Schlebusch, who said the company calculated about two million data points each time it rebalanced its portfolio.

"Machine learning equips fund managers with the tools to assess historical and present data, to help predict future risks and returns based on large volumes of data," he said.