Public.com launches AI agents for prediction markets — but only for events that can actually move your portfolio

Public.com is adding prediction markets to its platform, but with a deliberate constraint: no sports, no entertainment. Only macro-economic and company-specific events that could plausibly move a portfolio.

The distinction matters to founder Leif Abraham because it maps directly to who Public is building for. His framing is that Public serves the top quartile of US investors — people with money left over at the end of the month, managing retirement funds and life savings through the platform. Sports betting is entertainment finance, and he argues it belongs at arm's length from a serious portfolio.

The use case he returns to is one most active investors have lived: you're right about a company's quarterly performance, the stock still drops, and the loss comes from something orthogonal — a clumsy earnings call, an unexpected outlook, a macro event the same afternoon. Prediction markets let users isolate and trade a specific sub-KPI of a company rather than absorbing all the noise through an equity position.

“Today, we've launched AI agents for prediction markets. Our take on that is no sports, no entertainment, but essentially events that could impact your portfolio — events that move the market. / We had a guy come in who was with UBS Wealth Management, fired his wealth management team, moved $50,000,000 into a public account, now has AI agents managing his portfolio.”

AI agents + prediction markets

The more consequential piece isn't the markets themselves — it's how they feed Public's existing AI agent layer. Because the agents can act on real-time probability data, users can build conditional logic across asset classes. Abraham's working example: if the probability of a rate hike exceeds 70%, sell my treasuries. That kind of rules-based, cross-asset automation wasn't previously possible without prediction market signals as an input.

He's careful to position prediction markets as one signal, not the signal — a reasonable hedge given how loudly the forecasting-accuracy debate runs in this space.

Who's actually using the agents

The growth story Abraham tells is unexpected. The heaviest AI agent users on Public aren't retail beginners — they're sophisticated, high-net-worth accounts. He cites one user who left UBS Wealth Management, moved $50 million into a Public account, and now runs AI agents to execute covered call strategies for income generation. Tax-loss harvesting at scale is another example he gives: a simple, repeatable strategy that only meaningfully benefits accounts large enough for the tax savings to matter.

The implicit argument is that wealth management fees have historically been the price of access to systematic execution strategies. AI agents commoditize the execution, and the first people to notice are the ones who were paying the most for it.