BILL HARRIS LEFT PAYPAL YEARS AGO.. NOW HE'S LAUNCHING AN AI FINANCIAL ADVISER
Bill Harris was one of PayPal's early CEOs. Decades later, after Intuit, Personal Capital and several other fintech companies, he is trying to solve a problem that traditional financial advice never really solved: giving personalized advice to people who cannot afford a traditional adviser.
BILL HARRIS IS NOT NEW TO FINTECH
Bill Harris has spent roughly three decades building financial technology companies.
He was an early CEO of PayPal.
Then he became CEO of Intuit, the company behind products including TurboTax and QuickBooks.
Later, he founded Personal Capital, a digital wealth-management company.
Personal Capital eventually grew to about $23 billion in assets under management before being acquired by Empower in 2020 for approximately $1 billion.
So Evergreen.ai is not Harris's first attempt to change how people manage money.
It is his latest version of the same problem.
THE PROBLEM HE IS ATTACKING IS ACCESS
Traditional financial advice can be expensive.
And even when someone technically can afford advice, they may not have enough assets for an adviser to consider them an attractive client.
Harris's argument is that sophisticated financial planning has historically been concentrated among wealthier households.
Evergreen.ai is designed around the opposite idea:
Personalized advice → available to a much larger population
The company says its mission is to make financial advice available to every family rather than reserving sophisticated planning for wealthy investors.
BUT THIS IS NOT JUST CHATGPT FOR MONEY
This is probably the most interesting part of the product.
Evergreen.ai does use generative AI.
But the company says it deliberately does not want the language model to perform all of the financial mathematics itself.
Instead, the system combines:
AI language models
financial knowledge
user financial data
deterministic calculation engines
The language model handles communication and interpretation.
The calculation engine handles the numbers.
That distinction is extremely important for financial software.
WHY WOULD YOU SEPARATE THE AI FROM THE MATH?
Imagine asking an ordinary chatbot:
“Should I convert my traditional IRA to a Roth this year?”
A general-purpose AI can explain what a Roth conversion is.
But the answer depends on:
income.
tax bracket.
age.
retirement horizon.
existing retirement accounts.
future income.
state taxes.
conversion amount.
other deductions.
A generic answer can therefore sound intelligent while still being inappropriate for that specific person.
Evergreen's model is designed to connect the answer to the user's actual financial information and then use software specifically built to calculate the financial consequences.
THE USER CAN CONNECT THE FINANCIAL PICTURE
Evergreen.ai can work with linked:
bank accounts
brokerage accounts
credit accounts
and other financial information.
The idea is to create a more complete picture of the person rather than answering isolated questions.
That changes the interaction.
Instead of:
“What is a good retirement strategy?”
the user can potentially ask something closer to:
“Given my income, investments, savings and retirement goals, what should I consider?”
The software has more context.
And context is one of the biggest missing pieces in generic AI financial conversations.
THE PRODUCT CAN TRACK NET WORTH
Evergreen.ai includes net-worth tracking alongside planning tools.
That means the platform can combine:
cash
↓
investments
↓
retirement accounts
↓
credit
↓
other financial information
into a broader financial picture.
This is important commercially because the more of the financial picture a user puts into one platform, the more useful that platform can potentially become.
The product becomes less like a calculator.
It becomes a financial operating layer.
IT ALSO GOES AFTER TAX STRATEGY
Taxes are another area where generic financial advice can become complicated.
Evergreen includes tools and guidance around:
tax-loss harvesting
charitable giving
Roth conversions
equity compensation
and other tax-related strategies.
The company also specifically supports questions around stock options, RSUs, founder stock and QSBS.
That is a different customer problem from simply asking:
“How should I invest $500?”
The platform is trying to handle decisions where the answer depends heavily on the individual's broader financial situation.
HOMEOWNERSHIP BECOMES A FINANCIAL CALCULATION
Evergreen also includes mortgage calculations and rent-versus-buy analysis.
That sounds simple.
But buying a home changes several parts of a person's finances simultaneously.
Mortgage
↓
interest
↓
taxes
↓
cash flow
↓
investment opportunity cost
↓
long-term wealth
So the platform can potentially treat the home decision as part of the person's entire financial plan rather than as a standalone calculator.
RETIREMENT IS ANOTHER MAJOR USE CASE
The platform can model short-term cash flow and longer-term retirement projections.
That creates another important difference from a normal chatbot.
A chatbot can explain retirement.
A financial planning system can attempt to model:
income
→ savings
→ investments
→ withdrawals
→ retirement timeline
The output becomes a projection rather than just an explanation.
THE BUSINESS IS FREE — FOR NOW
Evergreen.ai launched its public beta on September 21, 2026.
Users who register during the beta receive access at no cost through January 1, 2028.
That is a very long free period.
And it makes the launch strategy interesting.
The company is not immediately trying to maximize subscription revenue.
It is trying to get people to:
Register
↓
connect financial accounts
↓
use the adviser
↓
build trust
↓
make Evergreen part of their financial routine
That creates a much more valuable relationship than simply selling an app download.
SO WHERE DOES THE MONEY COME FROM?
The free Evergreen.ai product sits alongside Evergreen Wealth Advisors, an SEC-registered investment adviser.
The company says additional services, including human investment-adviser services and investment management, are available through Evergreen Wealth Advisors for a fee.
That creates a potential funnel:
Free AI advice
↓
user understands financial situation
↓
user wants more sophisticated help
↓
human adviser / investment management
↓
paid relationship
The free AI product can therefore function as both a consumer service and an entry point into a broader wealth-management business.
THIS IS WHERE HARRIS'S PERSONAL CAPITAL EXPERIENCE MATTERS
Personal Capital already combined technology with wealth management.
It eventually accumulated about $23 billion in assets before its acquisition.
Evergreen is approaching the same broad problem with a much more powerful technological environment.
The difference is that the AI can potentially interact with the customer continuously.
Not:
Annual meeting with adviser
but:
Question → answer → calculation → follow-up → another question
The frequency of interaction can change dramatically.
AI COULD TURN FINANCIAL ADVICE INTO A 24/7 PRODUCT
A traditional adviser has appointments.
A software system does not.
Evergreen markets itself as available 24/7.
That matters because financial decisions do not only happen during office hours.
Someone might suddenly wonder:
Should I exercise these stock options?
Can I afford this house?
Should I increase my 401(k) contribution?
What happens if I retire at 60 instead of 65?
A conversational system can potentially answer immediately.
That changes the economics of access.
THE COMPANY IS ALSO TRYING TO SOLVE THE AI HALLUCINATION PROBLEM
Financial advice has a particular problem with AI.
A wrong answer about a movie is annoying.
A wrong answer about a tax calculation can cost real money.
Evergreen therefore says its calculation engine is deterministic.
In simple terms:
The AI can explain.
The software calculates.
That architecture is intended to make numerical outputs more consistent and verifiable.
The company still explicitly warns that AI can make mistakes.
PRIVACY BECOMES PART OF THE PRODUCT
There is another major challenge.
To personalize financial advice, the system needs financial information.
Evergreen says account linking is handled through Plaid, so bank credentials are not shared directly with the AI.
It also says user data is encrypted, is not sold to third parties and is not used to train public AI models.
This is more than a technical feature.
For a financial AI company, privacy is part of the product's ability to earn trust.
The proposition is essentially:
Give us more information → we can give you more personalized advice.
But that only works if users are comfortable giving the information.
HARRIS IS ALSO BUILDING FOR PEOPLE WHO ALREADY USE AI FOR MONEY QUESTIONS
Evergreen says consumers are increasingly turning to general-purpose AI tools for financial questions.
Harris's argument is that people are already asking AI about money.
The question is whether the AI actually has the information necessary to answer those questions properly.
That creates an interesting market shift.
The competition is not only:
Evergreen vs financial advisers.
It can also become:
specialized financial AI vs general-purpose AI.
THE DIFFERENCE IS THE DATA
Consider two systems.
GENERAL AI
Question
↓
general knowledge
↓
answer
EVERGREEN
Question
↓
user's accounts
↓
income
↓
investments
↓
goals
↓
tax situation
↓
financial calculations
↓
personalized answer
The second system potentially knows much more about the person.
That is its central competitive idea.
THE FINANCIAL ADVISER COULD BECOME SOFTWARE
This is the bigger business concept.
Historically, financial advice required:
human expertise
time
meetings
documents
calculations
portfolio software
AI can potentially compress some of those processes into software.
That does not eliminate human advisers.
But it can change what humans are needed for.
The software can handle repetitive analysis.
Humans can focus on complicated decisions, relationships and situations requiring judgment.
HARRIS HAS ALREADY BUILT SEVERAL VERSIONS OF THIS IDEA
His career creates an unusual sequence:
PayPal
→ payments
Intuit
→ financial software
Personal Capital
→ digital wealth management
Evergreen Wealth
→ technology-driven advisory
Evergreen.ai
→ AI-driven personalized financial guidance
The companies are different.
But the underlying theme is remarkably consistent:
use technology to put financial tools into the hands of more people.
THE FREE PRODUCT CAN CREATE A LARGE TOP OF FUNNEL
The economics could eventually look like this:
Millions of users
↓
free financial questions
↓
financial data connected
↓
personalized planning
↓
some users need advanced services
↓
investment management / human advice
The free product does not necessarily have to monetize every user directly.
It can create a large audience from which higher-value financial relationships emerge.
AND THE CUSTOMER'S FINANCIAL LIFE CHANGES
This creates another potential advantage.
A user might initially join because they want to understand retirement.
Later they need help with:
taxes
Then:
stock options
Then:
buying a home
Then:
college savings
Then:
investment allocation
The number of reasons to remain inside the platform increases.
The product becomes more valuable as the user's financial life becomes more complicated.
THE BUSINESS LESSON
The most interesting part of Evergreen.ai is not simply that an AI can answer financial questions.
It is that Harris is trying to combine three things that historically lived in separate businesses:
financial data
financial calculations
financial advice
If those three layers work together, the financial adviser becomes less like a person you meet once a year and more like software that sits continuously beside your financial life.
MAACAT PERSPECTIVE
Bill Harris left PayPal decades ago.
But the problem he is attacking has barely changed:
good financial knowledge exists — access to it is uneven.
His new strategy is different because AI can potentially make personalized analysis available at software scale.
The interesting business model is therefore not simply:
“AI gives financial advice.”
It is:
Free AI advice → financial data → personalized planning → deeper financial relationship.
The adviser may no longer begin with a meeting.
It may begin with a question typed into a box.
