Public.com says it has built the world's first agentic brokerage. The product page lets retail investors create AI agents that monitor markets, manage cash positions, and execute trades autonomously. Whether the label is earned or marketing depends on one question: does prompting an AI to execute a trading strategy constitute financial advice, or does it remain the user's decision dressed in new clothes?
The question is not rhetorical. It goes to the center of how US securities law treats AI tools that take actions on behalf of users — and the honest answer in early 2026 is that nobody knows for certain.
The regulatory map is not blank. The SEC's 2026 exam priorities list AI use as a focus area for registered investment advisers, and the SEC's Investor Advisory Committee issued disclosure recommendations in December 2025 as part of an ongoing examination of how AI tools interact with securities regulation. The existing framework — the Investment Advisers Act of 1940, Regulation S-P, broker-dealer suitability standards — applies to financial tools that either provide advice or direct investment behavior. What none of these rules clearly answer is whether an agent that executes a user-defined prompt is "directing" investment behavior in the regulatory sense.
The SEC under Gary Gensler proposed a predictive data analytics rule in 2023 specifically targeting technologies that "optimize for, predict, guide, forecast, or direct investment-related behaviors or outcomes." The rule was withdrawn earlier this year, leaving advisers to navigate existing frameworks. The withdrawal was not a deregulatory gesture — it was an acknowledgment that the existing rules did not cleanly map onto AI tools, and that new rules required more deliberation. That deliberation is still ongoing.
Public's own disclaimer reflects the ambiguity. "Outputs from Agentic Brokerage are provided for informational and illustrative purposes only, and should not be considered investment recommendations or advice," the product page states. "You are solely responsible for determining the suitability of any strategy and for verifying your instructions." This is legally correct as a matter of platform liability — the platform is not providing advice, it is executing user instructions. But it sidesteps the harder question: whether the friction removed by the agent changes the nature of the decision being made.
A human investor who decides to sell covered calls on FLYF because a newsletter recommended it is making their own decision. A human who sets an agent to sell covered calls on FLYF when implied volatility exceeds a threshold is also making their own decision — but the cognitive step between "define the rule" and "the trade executes" is compressed in a way that may or may not change behavior. If it changes behavior at scale — if retail investors begin using agents to execute strategies they would not have executed manually, or to monitor markets in ways they did not before — the systemic question becomes whether that shift in behavior creates risks that the existing suitability framework was not designed to address.
The mechanics of the product are real and the early access launch is genuine. Users write prompts, the system proposes a workflow, the user refines and activates it, and the agent monitors and executes within defined parameters. The example agents on the product page — trimming bank stocks on Fed rate cuts, sweeping cash above a threshold into bonds, placing a put if earnings call language crosses a frequency threshold — are concrete and the underlying execution infrastructure is conventional: Open to the Public Investing for brokerage services, Zero Hash for crypto. Several capabilities listed on the page, including shorting, tax-lot management, and trailing stop-losses, are marked coming soon rather than live.
The honest version of the story is this: Public has shipped a rules engine with a natural language interface to retail investors, positioned it as the world's first agentic brokerage, and placed the liability for strategy suitability entirely on the user. Whether that amounts to a new category of financial service or an existing one with better UX is a question the regulatory framework has not resolved. The product exists. The rules do not.
