SAN FRANCISCO, September 23 — Bill Harris wants to put a financial planner in your pocket, and he is giving it away for now. The former PayPal and Intuit chief executive rolled out Evergreen.ai on September 21, a consumer-facing application that uses artificial intelligence to build financial planning guidance around a user’s actual accounts, goals and tax situation. The product is in public beta.
Harris is positioning it as a way to make sophisticated planning available without the usual barriers of wealth, geography, or an appointment with an adviser. Anyone who registers during the public beta gets free access through January 1, 2028, with no credit card required.
Investors will note the pricing strategy: a giveaway window that runs well past a year, aimed at pulling users in before any revenue model kicks in.
A familiar name in fintech
Harris is not a newcomer pitching a concept. He previously led Intuit, served as CEO of PayPal, and founded Personal Capital, the digital wealth platform later acquired by Empower. Evergreen.ai is presented as the latest chapter in that career — software meant to help people understand, protect and grow their money.
On the product site, he argues that sound counsel has long existed and the missing piece has been access. The new app is his attempt to close that gap.
Unlike a generic chatbot, Evergreen.ai is designed to assemble a single picture of a household’s finances. Users can begin by asking questions in ordinary language, then add accounts — often through a secure Plaid connection — or enter remaining details themselves so answers rest on real balances rather than hypotheticals.
The company says the system covers planning, tax strategy, equity compensation, estate questions, retirement and income, investments, real estate and charitable giving, and that it is built to show how one choice can ripple through taxes, retirement timing and the rest of a plan.
The math is not the model
A notable design claim is that the numbers themselves are not generated by a large language model. Evergreen.ai says calculations come from tested, rules-based engines that use current tax and financial figures, so the same inputs should produce the same results. The conversational layer helps users explore scenarios — such as retiring earlier — while the underlying math is meant to stay consistent.
The company still warns that AI can err and that examples on the site are illustrative, not guarantees. Privacy is a central selling point.
Personal financial information is encrypted in transit and at rest, is not sold, and is not used to train public AI models, according to the firm. Sessions are described as isolated to the user’s account. Linking accounts is optional; people can start with questions alone, though connected data makes guidance more specific.
Evergreen.ai is a self-service tool. Using it does not make someone a client of Evergreen Wealth Advisors.
Paid human advice and investment management remain available separately through that affiliated advisory firm. Harris has also been building Evergreen Wealth, a registered investment adviser aimed at affluent clients, pairing fiduciary advisers with an AI engine and tax-aware portfolios. Evergreen.ai appears to extend related technology to a broader audience that may never hire a wealth manager.
In a July 2026 conversation at the New York Stock Exchange, he said many people already ask general-purpose AI about money, yet those tools often lack context and can hallucinate. A finance-specific system, he argued, should be available around the clock and cheaper than sitting down with an adviser. Whether the beta delivers on that promise will depend on accuracy, security, and how well the “whole picture” model works in practice.
For now, Harris is betting that planning software — not just account tracking — can be offered widely, and that domain-specific AI can do what generic chatbots cannot. The next test is the beta itself: whether users trust a machine with their balances, and whether the rules-based engine holds up when real financial lives get messy.
