Imagine telling an AI agent how much risk you’re willing to take, your retirement goals, and when your kids will start college — then letting it manage your portfolio while you sleep. That vision of agentic trading, where artificial intelligence doesn’t just recommend investments but actually carries them out, is moving from concept to reality. Brokerages, startups, and even retail investors are building AI agents that can oversee portfolios and automate investing tasks once handled by humans. “Effectively everybody has their own family office that is working 24/7 for them while they’re awake or sleeping,” said Devin Ryan, head of financial technology research at Citizens. “This isn’t 10 years away. This is coming in the next few years.”
Rather than trying to create fully autonomous trading systems overnight, many firms are taking a gradual approach. Startup Podium Markets AI is among those building artificial intelligence specifically for investing. Its assistant, Ivy, analyzes a customer’s portfolio across multiple brokerage accounts and generates recommendations based on the investor’s goals and risk tolerance — but stops short of acting on its own. “The AI informs, but the human decides,” said co-founder and CEO Dirk Mueller-Ingrand. Larger brokerages are moving in the same direction, with Robinhood introducing tools in May that allow third-party AI agents to connect with customer accounts, and Public developing its own AI agents that can automate investing workflows within the platform.
Retail investors have already spent the past three years testing what general purpose AI can do since ChatGPT burst into the mainstream in late 2022, using tools like ChatGPT and Anthropic’s Claude to summarize earnings reports, research companies, and generate stock ideas with mixed results. Obioha Okereke, a 29-year-old technology consultant in Georgia, built an agent using Claude to search for undervalued stocks and options opportunities but still reviewed every recommendation before placing a trade. Others have been less fortunate. Thomas Schlossmacher, a 31-year-old retail investor who tested a trading agent after seeing claims online that AI could uncover profitable market patterns, said he “was just losing money consistently” and now believes anyone relying on an automated system probably wants a professional involved.
That tension highlights one of the industry’s biggest challenges: teaching an AI agent what an investor actually means is far harder than teaching it to buy or sell a stock. An investor might tell an agent to grow their portfolio aggressively, but does that mean taking on more volatility, concentrating holdings, using options, or accepting a greater chance of loss? An AI can faithfully follow instructions and still produce an outcome the investor never intended. That’s why many firms are building guardrails before giving AI greater authority. Public requires users to review and approve an agent’s workflow before it carries out any tasks. “You still have the last word,” said Public co-founder Leif Abraham. Ryan estimates agentic finance could increase transaction volumes at least tenfold, predicting that by the end of next year, the majority of transaction activity on some platforms will be done by agents rather than humans themselves.
