How AI is changing your investing experience

What once required a portfolio manager or a team of analysts can increasingly be done with the help of artificial intelligence. But AI is not transforming retail investing in a single way. Some platforms use AI and deep learning to manage portfolios, others use conversational AI to guide investment decisions, while large language models are helping investors digest financial information faster. Understanding the differences between these approaches may be just as important as understanding the technology itself.

AI as a portfolio manager

Amwal AI was founded in 2023 with the goal of giving retail investors access to the same quality of research and portfolio construction tools available to institutional investors and wealthy clients, Founder Mohamed Ameen tells us. The platform allows users to invest in diversified portfolios that provide exposure to US-listed equities and other asset classes, while dynamically adjusting allocations based on changing market and economic conditions. “Most investors in our region are left choosing between low-yield bank products, static robo-advisor portfolios, or passive ETFs that simply rise and fall with the broader market,” he adds.

The company’s core argument is that many investment products rely on fixed asset allocations that do not adapt to changing economic conditions. Instead, Amwal AI combines traditional portfolio optimization techniques with AI-powered economic analysis. The platform continuously processes economic data from major economies around the world and compares current conditions with roughly 70 years of historical data to identify periods that resemble today's environment. “We’re not trying to predict anything, we’re trying to understand what the economy looks like today,” Ameen says.

AI, machine learning, and deep learning models are embedded throughout the investment process. Some models analyze economic data and historical market behavior, while others compare current conditions with similar periods in history to generate signals across more than 40 asset classes, including equities, fixed income, commodities, gold, oil, and cryptocurrency.

The company also uses AI within its risk management framework. Models trained on decades of options market data continuously monitor changes in options pricing curves and investor positioning every five minutes to assess whether market risk is increasing or decreasing. The goal is to identify shifts in risk appetite early and adjust portfolios before those risks fully materialize in broader markets.

While AI generates the portfolio allocation recommendations, the company still maintains a human review process before changes are implemented. However, Ameen says portfolio managers do not override the system or select investments based on personal views. “No person sits down and decides which assets we should invest in or not invest in,” he says. Human oversight is primarily used to ensure that model outputs remain reasonable and do not contain anomalies.

AI as a financial advisor

Some platforms use AI as a digital financial advisor rather than an investment engine. Thndr's recently developed Alpha is one example. The platform helps users build investment plans based on their goals and risk tolerance, explains the reasoning behind its recommendations, prepares buy and sell orders, and suggests portfolio adjustments when allocations drift away from their target weights.

Rather than making portfolio allocation decisions itself, Alpha acts as a conversational financial advisor. In effect, it seeks to replicate some of the functions traditionally provided by a financial advisor, but in a scalable digital format.

AI as a research assistant

Large language models such as Claude or ChatGPT represent a third category of AI applications used in investing. Rather than managing money or constructing portfolios, they can act as research assistants. Investors can use them to summarize annual reports, analyze earnings call transcripts, compare companies, explain financial concepts, identify risks, or highlight growth drivers hidden inside lengthy documents. Tasks that previously required hours of reading can often be completed in minutes.

Ameen argues, however, that research assistance should not be confused with portfolio management. While LLMs can help investors understand information more efficiently, he believes they are less suitable for building investment strategies because their outputs are non-deterministic, making them difficult to backtest and verify. “If you take the exact same input and give it to the model again five minutes later, you could get a different output,” he says.

Successful investing requires more than access to publicly available information. “If you want to build an edge, you either need data that is faster than everyone else's or data that is broader than everyone else's,” Ameen says. That is one reason why Amwal AI focuses on aggregating large datasets from multiple sources and testing its models across decades of historical market data.