Ep 70 | A Smarter Way To Approach AI For Your Finances
In this episode, Priya Malani unpacks why more access to AI-powered financial advice hasn't translated into more financial confidence. She breaks down the two core problems: AI that hallucinates regulations with total confidence, and AI that gives population-level answers to questions that are deeply personal. The real issue isn't that AI is wrong, it's that most people don't know the right follow-up question to ask. You walk out of the episode with a four-variable framework that pressure-tests AI financial advice so you can feel confident trusting it.
Takeaways:
More access to financial advice hasn't produced more financial confidence, because most people take the first AI answer at face value without knowing what's missing.
AI gives population-level answers to individual-level problems, and without the right follow-up, the advice could be technically accurate and completely wrong for your situation.
The four variables that most often change the right answer are time horizon, tax situation, liquidity needs, and income stability.
AI is a research assistant, not a financial advisor. An advisor knows your life. AI responds to what you put in front of it.
Follow Priya Malani:
LinkedIn | Instagram | Youtube | Stash Wealth
The Stuff Our Lawyers Want Us to Say:
Stash Wealth is a Registered Investment Advisor. Content presented is for informational and educational purposes only and is not intended to make an offer or solicitation for any specific securities product, service, or strategy. Consult with a qualified investment adviser (that's us) before implementing any strategy. Investing involves risk, including the loss of principal. Past performance does not guarantee future results. There…we said it.
Transcription
Why More AI Access Doesn't Mean More Financial Confidence
That's the irony. The questions people most urgently want answered are exactly the questions where AI is most likely to give you a confident answer. That's not accurate. Who the am I to tell you what to do with your money? My name is Priya Malani, currently managing millions of hard working dollars. Enough for a play. Let's talk. Welcome to the F word smart money, no? Hey guys, welcome to the F word.
AI is changing a lot of things, and one of them is access to financial advice. A 2025 report from Intuit Credit Karma found that 66% of Americans have already used generative AI tools like ChatGPT for financial advice. Among Gen Z and millennials, that number climbs to 82%. And a separate study from FNBO, First National Bank of Omaha found that about half of Americans trust AI when it comes to their finances. Those numbers make sense. AI is fast. It's free. It's available at 11:00 PM when you're anxious about your money and can't sleep. But here's what's interesting: as access to free financial advice and personal finance apps has increased over the past several years, our overall financial confidence hasn't kept pace. In fact, the same study from FNBO found that fewer than 2/3 of Americans feel confident in their long term financial plan, and one in three say they feel behind their peers financially. So there's more access to financial advice than ever, but not more financial security.
That gap is worth discussing. Here's what I think's driving part of it. People are using AI for financial questions, but most are taking the first answer at face value without realizing whether that answer actually applies to their situation. That's what today's episode is about, not whether you should use AI for finances. I think you should, especially in the beginning, but I want to talk about how to use it the way a smart professional would as a starting point that you pressure test, not a conclusion that you can simply act on.
Let's get into it. So let's start by thinking about how you use AI anywhere else, especially at a place like work. If you're a strong writer and you ask AI to write you an email, you don't just copy, paste and hit send. You read it. You rewrite a few sentences, you flag where it missed the tone or buried the point, or even got the facts slightly off. Your expertise is doing the quality check. Same thing if you're in marketing or law or software engineering. Professionals who use AI well treat it as a first draft, a launching pad. But then they apply their own domain knowledge to push, refine and challenge until the output is actually right.
Now think about how most people ask AI financial questions. They type something in, get a confident, well structured answer and act on it. They don't have the financial background to know which part of the answer to challenge or what question to ask to sharpen it. So the first response becomes the final one. That gap between the confident sounding answer that AI gives and the right answer for your specific situation is where the money gets lost.
The problem isn't that AI is bad. The problem is that financial advice is deeply personal and AI is giving you population level answers to individual level problems. It knows what's generally true. It doesn't know what's true for you unless you tell it, and most people don't know what to tell it.
Here's what I'm hoping you'll take away from this. Think of AI's first answer as the floor. Your job is to get to the ceiling. That means asking the follow up questions that give AI the context it actually needs to help you specifically.
Beware: AI Can Fabricate Financial Facts and Outdated Information
Before we get into the how, I want to spend a few minutes on another problem that goes beyond incomplete advice. Because AI doesn't just give you advice that's missing context. Sometimes it gives you advice that's simply factually incorrect. And the dangerous part is that it sounds exactly like the correct answer.
This phenomenon is called hallucination. It's when AI generates information that sounds plausible, authoritative, even well sourced, and it's completely made up. Here's a real example. MIT Professor Andrew Lo, one of the top researchers in financial engineering in the country, used ChatGPT for financial research and found that it fabricated the authors' names for a paper it cited to back up its own recommendations. The paper was real, but the cited authors were not. The AI invented credible sounding names to fill in the gap in its knowledge and presented it as fact.
That's the thing about hallucination. Unless you know better, you don't know any better. It doesn't say I'm not sure about this part. It delivers invented information with the same confidence that it delivers accurate information. In personal finance this has real stakes. Regulations change, tax laws change, contribution limits change, interest rates change, and AI's knowledge has a cutoff date. If you ask about current 401K limits, or whether a particular tax strategy is valid, or what rate a certain type of account is currently paying, you may get an answer that was accurate 18 months ago and isn't anymore. The CEO of a financial AI company said that if you ask AI about a lesser known company or financial instrument, it may simply fabricate an answer to give you something rather than admitting that it doesn't know.
Where AI tends to struggle most is with timely, specific questions. Current rates, current regulations, specific investment vehicles, contribution limits that may have changed, and yes, even in my own experience, straightforward math. It's far more reliable on conceptual, evergreen questions than it is on anything requiring up-to-date facts.
The answer can be technically correct and still be completely wrong for your situation.
Why AI's 'Correct' Answers Can Be Wrong for Your Situation
Let's look at a couple Q&As together so you can see the pattern.
Question for AI: should I max out my 401K? AI says yes, great tax savings, great long term move, and it is generally, but if you're planning to buy a home in the next two to three years, you may have just locked your down payment behind an early withdrawal penalty. The follow up AI should have asked: are you saving for anything in the near term? That changes the answer completely.
Question for AI: should I invest in index funds? AI says yes, low cost, diversified, one of the best long term strategies out there. All true, but if you need that money in 18 months you have no business being in equities at all. A market correction at the wrong moment turns a sound long term strategy into a short term loss that you can't recover from in time. The follow up AI should have asked: when do you actually need this money?
Question for AI: should I do a Roth IRA? AI says tax free growth is powerful, and it is. But if you're currently in a high income bracket and expect to be in a lower one in the future, you're paying taxes now at your highest rate and miss the ability to pay them later at a lower rate. A Traditional IRA might put significantly more money in your pocket over time, not to mention increased cash flow if cash flow is tight. The follow up AI should ask: what's your tax bracket today versus what do you expect in retirement? Also, do you have debt on the books that you're trying to pay off?
Question for AI: should I buy or rent? AI says buying builds equity and renting is money you never get back, so buying wins long term. Except if you're not staying put for at least five to seven years, the closing costs, transaction fees, property taxes, and the fact that your early mortgage payments are almost entirely interest can mean buying actually costs you more than renting would have. The follow up AI should ask: how confident are you that you're staying put for a while?
Question for AI: should I refinance my mortgage? AI says yes if you can get a lower rate. Sounds straightforward, but refinancing comes with closing costs, typically 2 to 5% on the loan, and if you sell or move before you've broken even on those costs, you've lost money on the deal even with the better rate. The follow up AI should ask: how long do you plan to stay in the home?
Question for AI: should I take the job with the higher salary? AI says more income improves your financial picture, but if the new job comes with worse health benefits, no 401K match, a higher cost of living city, and a bump in your tax bracket, the headline number can be almost meaningless. Total compensation and net take home are what actually matter. The follow up AI should ask: what does the full package look like after taxes and benefits?
Question for AI: should I open multiple credit cards for the rewards? AI says yeah, you can maximize points, cash back, and benefits across all the cards. Mathematically that's true, but the optimal reward strategy requires tracking multiple statements, payment dates, and spending categories. Miss a payment on one and the fees and interest wipe out months of rewards. The follow up AI should ask: are you someone who will genuinely manage that complexity every month without fail?
6 questions, 6 confident, reasonable AI answers, 6 situations where one follow up question changes the advice entirely. That's the pattern, and the reason most people never get around to the follow up question is because the first answer sounds complete. It doesn't feel like there's a follow up question to be asked.
The Four-Variable Framework to Confidently Trust AI Financial Advice
So here's my system for those of you using AI for financial advice. After you get any financial answer from AI, run it through 4 questions before you act on it. There are 4 variables that most commonly change the right answer for you.
First up, time horizon. When do you need this money or by when does this decision need to pay off? Advice for a 20 year horizon and advice for a two year horizon are often the exact opposite.
Next, tax situation. Do you know your current tax bracket? Your expected tax bracket in the future? Your state tax situation? AI will give you a generic tax answer, but your tax answer may be different.
3rd, liquidity needs. Can you afford to have this money locked up? The optimal move on paper becomes a really expensive trap if you need access and can't get it.
And finally, income stability. Is your income salaried and predictable or variable? This changes emergency fund size, insurance recommendations and risk tolerance calculations significantly.
After AI answers your financial question, ask yourself which of these four did it ignore? Then go back and ask AI that specific follow up. You'll be surprised how often the advice materially changes. And one more thing, if the answer involves specific numbers, a current rate, a current legal limit, a regulatory rule, verify it independently. Don't trust the number. Look it up. AI is far more reliable on principles than on current facts. I hope that helps.
Best Bite
I had lunch this past weekend at La Coucou in Soho and I ordered the tartare de boeuf. Apologize for my French. It was topped with caviar and a quail egg. It was by far the best tartare I have ever had and I do order it very often. I was there for lunch. I presume it's on the dinner menu. Otherwise go for lunch. Tartare de boeuf from La Coucou. That was my most recent best bite.
Before I let you go, here's the thing I want you to remember. Use AI as a research assistant, not a financial advisor. An advisor knows your life. A research assistant responds to what you put in front of it. The quality of what comes out is directly tied to the quality of what you put in. The people who get the most from AI for financial planning aren't the ones who ask and then accept whatever it says. They're the ones who ask, read the answer skeptically, identify what's missing, and then probe again. They treat every first response the way a smart professional treats a first draft. Useful, directional, but not ready to ship. It's kind of like your intern.
Let me be clear, you do not need a financial degree to do this. Just remember my 4 point framework for following up. What's my time horizon? What's my tax situation? Do I need this money to stay liquid and is my income variable or stable? One of those will almost always be the thing the first answer missed. If this episode gave you a smarter way to approach AI for your finances, that's great news. And if you have a minute to help someone else find the show, the best thing you can do is leave a review or share it with a friend who needs to hear this. As always, like, subscribe or follow wherever you're listening. I appreciate it. All right, that's it for today. See you next time.
Subscribe, Review, and Important Legal Disclaimer
Thanks for listening to the F word with Priya Malani. If you like what you heard, hit subscribe wherever you're listening and leave us a review while you're at it. We're approval junkies. Don't forget, you can find a ton of great resources, content, courses and other freebies at stashwealth.com. Now for the capital F stuff our lawyers want us to say: Stash Wealth is a registered investment advisor. Content presented is for informational and educational purposes only and is not intended to make an offer or solicitation for any specific security product, service or strategy. Consult with a qualified investment advisor — that's us — before implementing any strategy. Investing involves risk, including.
THE STUFF OUR LAWYERS WANT US TO SAY: Stash Wealth is a Registered Investment Advisor. Content presented is for informational and educational purposes only and is not intended to make an offer or solicitation for any specific securities product, service, or strategy. Consult with a qualified investment adviser (that's us) before implementing any strategy. Investing involves risk, including the loss of principal. Past performance does not guarantee future results. There…we said it.
