5 AI Prompts That Will Change How You Manage Money (And 3 Things It Still Gets Dead Wrong)

AI doesn't do financial planning. You do financial planning. The AI just makes it more accessible, more interactive, and potentially a lot cheaper. But the quality of the output is a direct function of the quality of your input. Garbage in, garbage out. And in financial planning, garbage can be measured in the size of your retirement account. Hello friends, this is Tyler Gardner welcoming you to another episode of your money guide on the side, where it is my job to simplify what seems complex, add new ones to what seems simple, and learn from and alongside some of the brightest minds in money, finance, and investing. So let's get started and get you one step closer to where you need to be. Welcome back to your money guide on the side. I'm Tyler Gardner and today we're doing something that I've been genuinely looking forward to for a while. Because the number of questions I've gotten on this topic has officially reached what I call the critical mass level, which is the level where I stop answering individual messages and make a whole episode instead. So congratulations collective inbox you win. Today's episode artificial intelligence as your financial advisor. I did an episode on this about a year ago at this point and it got a lot of positive responses, but even just over the past year, we've gone from AI might be helpful and cool to holy crap, this thing is pure insanity and if you're not using it in some way daily to help you optimize your life, well, I think you're truly missing out. Additionally, I think this might be the most important episode I've recorded since the tools that you're disposal now could quite literally change your life as they certainly have mine. But before anyone goes and writes in the comments and someone always does, Tyler, AI can't replace a human advisor. I want to be very clear up front. I know that's actually the point of the episode. We're going to talk about what AI can do, which is quite a lot, what it cannot do, which is also quite a lot. And we're going to do the thing we always do on this show, which is look at some real data and use that data to make better decisions instead of just guessing our way to and through retirement. I know, wild concept. But first, and I mean this every single time I say it, if this show has been even remotely useful to you, if you've learned something, made a better financial decision, or even just felt slightly less like you're winging your entire financial life, please consider leaving a review on Apple Podcasts or Spotify or wherever you listen. Your review helps new people find the show, which affects whether more people learn this stuff, which is genuinely the entire point of the entire endeavor. We're going to start section one with what I'll call how to actually use AI as a financial guide. Let me start with a number that I think about a lot, and I've shared with you a lot. According to Dalbar, that's the research firm that's been tracking investor behavior since 1994. The average equity investor has historically underperformed the S&P 500 by somewhere between 1.5 and 4% points per year, every year for 30 years. Now, just to be clear about what that means in practice, if the market returned 10% and you returned 6.5%, you didn't lose money. You just made significantly less of it than you would have had you just stuck the money in an index fund, and compounded over a 30-year career. That gap doesn't look like a minor rounding error. It looks like retiring with about half the money you should have had. Now, I'm not telling you to just put all your money in a stock index, as that's not always going to be right for you, and we'll explore exactly how you can figure out what is right for you throughout this episode. But you do need to always be looking at what are called benchmarks, which just means how the market indexes are doing on their own and see how your portfolio is doing relative to those indexes. Now, why do we underperform on average a basic index? Well, according to Vanguard's research, and they've studied this stuff exhaustively bless their souls, the single biggest source of value a financial advisor provides isn't picking stocks. So stop asking a prospective advisor if they can beat the market. That's not what they're there for, especially in this era with the abundance of information we all have. Their value primarily comes from what they call behavioral coaching, specifically stopping you from doing something catastrophically stupid when the market drops 30% and your lizard brain is screaming at you to sell everything, buy gold, and renovate the basement so it serves as your future bunker. Vanguard estimates this behavioral coaching is worth about 1.5% per year in added returns, which is remarkable because what they're essentially saying is a good advisor's primary value is talking you out of talking to yourself. So let's enter this episode, agreeing that most financial advisors these days are saying their added value in theory and practice is to be your behavioral coach. Okay, now here's where AI enters the picture because it turns out, and I've spent a lot of time talking about this and testing this, a well prompted AI can do a genuinely impressive version of exactly that type of behavioral coaching, not necessarily all of it, but far more than you might expect. Here's the framework I want to give you today because using AI for financial guidance is not like asking what temperature we should bake chicken at. The output is only and forever as good as the input, and most people, no offense, are asking terrible questions to AI and then deciding AI is useless, which is like giving someone a violin for the first time, deciding their terrible at music and blaming the violin. The prompts matter enormously, so I'm going to give you the specific prompts that actually work. You can write them down, pause the episode, or just subscribe to the newsletter at TylerGardner.com, do whatever you need to do. These are real, and if a human can get you an additional 1.5% return by being your coach, just imagine what a 24/7 access language model can do for you if prompted well. Step one, we're going to build your financial profile. The first thing you need to understand about AI financial conversations is that the model has no idea who you are. Even from one session to the next, it can forget every input you gave it unless prompted otherwise. It doesn't know your age, your income, your debt, your goals, your portfolio, your risk tolerance, or the fact that you panic checked your brokerage account 17 times during the 2022 drawdown. You have to tell it, and then you have to tell it again. Always. Now, I'll come back to the privacy concerns around this in a bit, and they are real. But for now, let's talk about how to create a financial profile effectively. The single most important thing you can do before any AI financial conversation is give it a complete financial snapshot, and I do mean complete. Here is prompt number one, the financial snapshot. You're going to actually enter this exact text. And again, this is in the newsletter this week, and in the show notes, if you want the actual language. I want you to act as a knowledgeable, objective financial guide, not a licensed advisor to help me think through my financial situation. I'm going to give you my full financial picture, and I want you to ask me follow-up questions until you have everything you need to give me useful, specific guidance. Here's where I am. Age, household income, monthly take home pay, current monthly expenses, break that down into housing food, transport subscriptions, emergency fund, what types of debt, what types of current retirement accounts, 401k, IRA, pension, any other investments, taxable brokerage, real estate, do you even have an employer 401k match? Then time horizon to retirement or your respective goal, and then what you think your risk tolerance actually is, conservative, moderate, aggressive, then your biggest financial fear, let's be honest, and then your financial goal for the next 12 months. Finally, you would close the prompt by saying, "Please review what I've shared. Ask me anything that would help you give better guidance and then tell me the three most important things I should do first." The reason this prompt works is because it forces you to articulate your situation in full, which most people have never done in one place, and it tells the AI to ask follow-up questions rather than just give generic advice. That follow-up loop is where the real value is. Step two, understanding your actual risk tolerance. Here's a fact that should humble all of us. A Morgan Stanley survey found that roughly 70% of investors don't actually know their own risk tolerance. tolerance. Blondley, I'm surprised it wasn't higher. They think they do. They'll tell you they're moderate. And then a correction happens and they discover they are, in fact, extremely not moderate. Risk tolerance has two components that most people can flate. There's your ability to take risk, which is mathematical, based on time horizon and financial situation. And there's your willingness to take risk, which is psychological, based on how you actually behave when things go wrong. Both matter, and they're often miles apart. Here's a prompt that actually stress tests this in a useful way. This episode is brought to you by Gelt. Here's a question I bet you've never asked yourself. What's the difference between a tax preparer and a tax strategist? A tax preparer shows up once a year, collects your documents, fills out forms, and tells you what you owe. That's it, transaction complete. A tax strategist, they're calling you in October saying, hey, you're about to cross into a higher tax bracket. Let's talk about accelerating some expenses. They're texting you in December. Don't take that distribution yet. Let's run the numbers first. If you're a small business owner or a high earner, that distinction is worth tens of thousands of dollars a year. Gelt isn't your dad's CPA firm. They're proactive. They reach out throughout the year with actual strategy, not just compliance. Should you buy that equipment before year end? Should you pay yourself a bonus or leave it in the business? What's your estimated payment schedule look like to avoid penalties? And here's what surprised me. Their platform doesn't feel like tax software from 1997. It's actually organized, intuitive, and built for people who run businesses. Not accountants who love pivot tables. So if you've been avoiding finding a real tax partner because you assume it's expensive or complicated, stop. Gelt can save you money and simplify your life. And they'll give you a free consultation to see if it's the right fit for you. Head to joingelt.com/tyler. That's j-o-i-n-g-e-l-t.com/tyler. This is prompt number two, finding your real risk tolerance assessment. You're going to enter. I want to figure out my actual risk tolerance. Not what I think it is, but what it probably should be based on both my financial situation and my psychology. First, here's my financial situation. This is where you would paste your prompt number one or the summary. Now I want you to walk me through three scenarios and I'll tell you honestly how I'd react. Scenario A, the market drops 15% over three months. My $100 portfolios now worth $85,000. What would I want to do? Scenario B, the market drops 35%, similar to 2022, 2008. My $100,000 is now worth $65,000. It's been down for eight months and no one knows when it will recover. Scenario C, a single stock I own drops 60% after bad earnings. My entire position that was worth $20,000 is now only worth $8,000. What would I want to do based on my answers to these scenarios and my financial situation? What is my actual risk tolerance and what asset allocation does that suggest? Please also tell me if my stated risk tolerance and my behavioral risk tolerance appear to be different and what that means for how I should invest. This prompt works because it uses scenarios instead of abstract questions. Are you comfortable with risk? Is a meaningless question? Of course, everyone says yes in a bull market, but your portfolio is down 35% and has been for eight months. That question has a different answer. That's the answer that actually matters for how you should be invested. Step three, build an actual allocation strategy. Okay, once you have your financial profile and your real risk tolerance, you can ask AI to help you build an allocation strategy. And here is where it genuinely shines because it can explain the logic behind different approaches in plain English. You specific fund examples and show you how to think about this across different account types. Now, I want to flag this very clearly because I'm going to revisit it in section two of this episode. You need to verify any specific fund names, expense ratios and account limits it gives you against current sources. AI's knowledge can have a cutoff date and a fund that existed at.03% expense ratio last year may have changed. Fund names get merged, limits get adjusted for inflation. The framework it will give you will most likely be solid, but the specific numbers you must cross check. So with that caveat, in the air, like a slightly awkward balloon hanging at a birthday party, here's the prompt you would enter here. Prompt number three, allocation strategy with fund examples, you're going to enter the following. Based on my financial profile, this would be from prompt number one, and my risk tolerance. This would be from prompt number two. Please help me build a simple diversified investment allocation strategy, specifically what percentage allocation across major asset classes, stocks, international stocks, bonds, alternatives, cash, make sense for me, and why? Two, for each asset class, give me two or three specific low cost index fund examples I could use. Please include the fund name, ticker, and approximate expense ratio, and note that I should verify these are current before buying. Three, how should this allocation differ across my different count types? For example, what goes in my 401k versus my Roth IRA versus my taxable brokerage, and why? Four, what's the one biggest mistake someone in my situation typically makes with their allocation, and how might I avoid it? And five, how often should I rebalance, and what should trigger me to revisit this strategy entirely? Please explain the reasoning behind each recommendation in plain English, I want to understand the logic, not just the instructions. The reason I love this prompt specifically is point three, the asset location question. This is one of the most underrated concepts in personal finance. Putting the right investments in the right account type can add meaningful returns without changing your allocation at all. It's free money in the most literal sense, and most people have no idea they're leaving it on the table. Step four, stress test with worst case scenarios and Monte Carlo simulations. Let's talk about Monte Carlo simulations for a second because this is something professional financial planners use routinely, and it sounds intimidating, but it's actually a very simple concept once you understand it. A Monte Carlo simulation runs your financial plan through thousands of randomly generated market scenarios, not one version of the future, thousands of them. Some of those scenarios are great, some are terrible, some are mediocre, and at the end, you get a probability that reads something like this. Given your current savings rate, spending plan and investment allocation, you have an 87% chance that you won't run out of money before age 90. That's somewhat useful. That's a number you can actually do something with, and here's the thing, AI can walk you through the logic of this even if it can't run the actual simulation, but tools like projection lab and fire calc, or even the retirement calculators on Vanguard site can run through the actual numbers, and then AI can help you understand what those numbers mean and what levers to pull. So here is prompt number four for stress testing and worst case scenario planning. You're going to say the following. I want to stress test my retirement plan against realistic worst case scenarios. Here's my situation. Paste your prompt from number one. Please walk me through the following. One, what does a 2008 style sequence of return scenario look like for me? As in if I retire in X years and the market drops 40% in my first year of retirement, how does that affect my portfolio longevity versus a scenario where the drop happens 10 years before I retire? Two, what is a Monte Carlo simulation? And if you had to estimate my probability of a successful retirement, i.e. not running out of money by age 90, based on my numbers, what would be your rough estimate and what assumptions are you making? Three, what are the three most dangerous assumptions in my current retirement plan? The ones that if wrong would cause the most damage. Four, what's the minimum savings rate I would need to feel reasonably confident about retirement given my situation? And five, if I wanted to build in a margin of safety, a buffer against bad luck, what would you suggest I adjust first? Savings rate, retirement date, expected spending, or investment allocation? the sequence of return. that's genuinely important and genuinely underappreciated. Two people can have the exact same average market returns over their lifetime and end up with wildly different outcomes, depending entirely on when the bad years happen. Retiring into a bear market is devastating in a way that a bear market 10 years before retirement simply is not. This is worth understanding deeply before you get there. This week's episode is brought to you by Fabric, a friend of mine just had his second kid. And when I asked if he had term life insurance, he gave me the exact look you'd expect. The one that says, "I know, I know." Well, clearly meaning, I have not thought about this once. He's not irresponsible. He's just busy. And this is the kind of thing that lives permanently on the to-do list next to "clean out the refrigerator." But if anyone depends on your income, term life insurance isn't optional. It's the financial safety net that keeps everything intact if you're not around. That's why it sits at step four in my financial order of operations before contributing to an emergency fund as this truly is your emergency fund. Fabric by Gerber Life lets you apply in about 10 minutes online. No health exam, no phone calls, no laminated brochures. A million dollars in coverage can run less than a dollar a day, especially if you start young. And if you're thinking, "I have coverage through work." You might want to check the actual number. Most employer plans cover one to two times your salary and disappear the moment you leave. 10 minutes go to meetfabric.com/tiler and cross it off the list. That's meetfabric.com/tiler. Policy is issued by Western Southern Life Assurance Company, not available in certain states, prices subject to underwriting and health questions. Step 5. AI as your behavioral coach. And now my favorite part. This is the use case that I think AI is genuinely, maybe surprisingly excellent at. And it's the one that has the biggest dollar value attached to it, which if you remember from the beginning, Vanguard put at about 1.5% per year. It's ability to talk you out of bad decisions. The market drops, your portfolio is down 20%. Your neighbor sold everything last week and is looking very smug about it. CNBC has a countdown clock to some kind of economic catastrophe. And you're sitting there with your hand hovering over the cell button, texting your spouse something that begins with, "Hey, I've been thinking, this is the moment, and this is exactly when you should open an AI and type the following. Prompt number five, your behavioral guard rail. You're going to say, "I need you to talk me through a financial decision I'm about to make, and I want you to be direct with me, even if that's uncomfortable. Here's my situation. You're going to describe your portfolio, your allocation, your time horizon, your financial goals. Here's what I'm thinking about doing. Describe the thing you're about to do, sell, buy, change allocation, move to cash, etc. Here's why I want to do it. Be honest in this part. Is it fear, a tip from your buddy, the news, a gut feeling? I want you to do three things. One, tell me what the historical data says about making this kind of move at this kind of moment. Two, tell me what the argument for doing this is. Because I want to hear the best case for my own decision. Three, then tell me what you actually think I should do and why don't hedge. Give me your honest read. After you do all three, ask me, what would have to be true for this to be a good decision? That last prompt, the last question, what would have to be true for this to be a good decision? Is one of the most useful thinking tools I know? Because it forces you to articulate the actual premise of your decision. And usually, when you're articulated, you immediately realize the premise is, I'm scared. Dressed up in economic language, which is not historically a great basis for portfolio decisions. A 2020 study published in the Journal of Financial Planning found that investors who sought some form of guidance or coaching before making major portfolio changes during market volatility made significantly better decisions than those who acted immediately on impulse. Now, that study was about human advisors, but the behavioral mechanism is the same. You have to externalize the decision before you make it. AI can serve that function at 2am when your advisor is asleep and the market is theoretically open in Tokyo. Which brings me to section three. These are the concerns and they are real. Okay, I promised you balance, and here is said balance. I genuinely believe AI is useful for the things we just talked about. I use it myself, I'll say that clearly, but there are real significant concerns with using AI as your financial guide and I want to be honest about all three of them. Because ignoring them is the financial equivalent of reading only the return projections in a mutual fund prospectus and skipping the risk section, which statistically describes most mutual fund investors, so maybe that's not a strong warning, but I'm going to try anyway. Concern number one. You're giving a language model your entire financial snapshot. I know, I just told you to do this, but I want you to think for a moment about what's in these prompts I just gave you. Your age, your income, your debt, type, amount, interest rate, your account balances, your retirement savings, your financial fears, and where your money is. And I want you to think about where that information goes. Here's what I can tell you. The major AI providers andthropic, open AI, Google, have different policies about how they use conversational data and those policies change frequently. Some of them use conversations to train future models by default. Some don't. Some offer opt out. Some offer paid tiers with stronger data protections. You have to go read the current policy for whatever tool you're using, because I cannot tell you what it says today, and even if I could, it most likely will have changed by the time you hear this. But what I can tell you is this. The people who used to email me looking for financial advice, would share with me their entire financial situations without ever having connected with me once. This is not great judgment about what to put into writing on the internet to a random influencer walking through the woods of Vermont. I have received, and I kid you not, people's full account numbers, social security numbers, screen shots of brokerage statements, complete tax returns, and these are people who thought they were just asking me for a basic advice without ever having met me. I'm a person with a podcast. I'm not a financial institution. I have no regulatory obligation to protect that information. Obviously, I do protect it. I just delete it, because I'm human being with a conscience. But the point is, you need to think before you share, and if you're putting what you sent to me into AI, well, I just remember this is not some highly protected database. It's a very, very public information with ever changing rules. So at a minimum, don't ever, ever, ever, ever, ever, ever, use real account numbers. Don't use your social. Don't take screenshots. Describe your situation in dollar amounts and percentages without using any identifying information. You don't even need to use the actual names of the companies you work with. You can use the enterprise or privacy-focused tier of whatever tool you're using if it's available, but you also need to understand that private and secure are not the same word. So if the information you're about to type into AI would make you even remotely nervous to see on the front page of your local newspaper, just pause and reconsider what level of detail you actually need to provide. Concern number two. Garbage in, garbage out, and the garbage can be dangerous. This one is more structural, and it applies to every use of AI, not just financial guidance. The prompts I gave you earlier in this episode took me time to develop. They're specific. They're structured. They force the AI to reason in a particular direction. The guidance you get from a well-constructed prompt, and the guidance you get from, hey, what should I invest in, are not in the same universe of usefulness, not even close. And here's why that matters. Vague prompts produce plausible sounding generic advice, and plausible sounding generic advice is not just unhelpful. It can be actively harmful because it sounds authoritative. It comes back in clean, confident pros with bullet points and specific sounding percentages. And if you don't know enough to evaluate whether it's correct, you might just act on it. Here's a real example of how this does go wrong. Someone asks an AI, "Should I put all my savings into a target date fund?" And the AI responds, "Target date funds are excellent for retirement savings. They automatically rebalance, they're diversified, and here's how they work." All of that is true. But what the AI doesn't know because the person didn't say it is that their 28 years old have $80,000 in credit card debt at 24% interest, no emergency fund, and is planning to buy a house in two years. In that context, put everything in a target date fund is genuinely terrible advice. But the AI gave it confidently and correctly for the question that was asked. The question that was asked was the wrong question and that's on the human that's garbage in. This is why I've spent most of section two on how to prompt correctly. Because an AI financial guide is only as good as your ability to ask the right questions in the right way. And if you don't know what the right questions are, which is reasonable because financial planning is genuinely kind of complicated, you may be getting advice that is technically responsive to your prompt and practically useless or worse. Now if you take no other prompt from this episode because I know those were long prompts, take this one. Hey Claude or hey chat GPT, please act as a financial expert and tell me what to ask you specifically so I can learn more about myself as an investor. If nothing else, prompt it to prompt you. I'll say that again. Prompt it to prompt you. Be specific and if the advice sounds too simple, it probably is. Concern number three, the hallucination problem and why it's particularly spicy in finance. Large language models hallucinate. This is not a bug that's going to get fully patched. It is a structural feature of how these systems work. They generate plausible text. Sometimes plausible text is accurate. Sometimes it isn't. And the model does not always know which is which, which is why it doesn't always tell you. In most contexts, hallucination is annoying. You ask for a book recommendation and the model invents a book that doesn't exist. That's inconvenient. But in financial context, hallucination can be expensive. Here are specific things in AI might get entirely wrong that you really, really do not want to be wrong about. 401k contribution limits. These are adjusted annually by the IRS. An AI trained on data from a prior year will give you last year's limit as if it might be current. Same goes for IRA contribution limits. Fund names and tickers. Funds can get merged, renamed, liquidated. A ticker that was a Vanguard fund two years ago might be something else today. Never buy a fund based solely on an AI recommendation without confirming the current ticker, expense ratio, and what it actually holds. Tax rules. The tax code changes. Roth conversion rules, capital gains thresholds, backdoor Roth eligibility. All of these can change with legislation and an AI may be working from outdated information. Interest rates and economic data. AI has a knowledge cut off. Any rate environment in Flation Figure, economic condition it describes may be months or years out of date. So here's the rule I suggest. Use AI to understand frameworks and concepts. And then use current primary sources. IRS.gov, Vanguard, Fidelity, Morningstar to verify any specific number before you act on it. The AI is your strategy consultant. The IRS website is your accountant. You need both. I'll put it another way. AI is an extraordinary map, but the road may have changed since the map was printed, so you still have to look out the freaking window. Section 4. My overall take. Where does this leave us? Well, here's what I actually think about AI as a financial tool. And I want to be precise because I think both of the evangelists and the dismissers are getting this part wrong. Can it be useful? 100% you better believe it and if you're not using it to complement your financial planning, you need to be. AI can talk you through why you shouldn't sell. Right now, tonight, when the market is down and you're spiraling, it can walk you through what happened in '08, 2020, 2022, every major correction in modern history, and it can remind you what people who sold the bottom got in return for their conviction, a locked in loss and the experience of watching the market recover without them. That's not nothing. That's genuinely valuable. It can put current economic events in global perspective. When something scary happens, like a war, an election, a banking crisis, or a global pandemic, AI can give you immediate context. Here's what happened to markets in similar historical situations. Here's the range of outcomes. Here's what the data says about the correlation between scary headlines and long-term equity returns, which is, for the record, shockingly low, scary headlines are not a reliable predictor of long-term market performance, but your amygdala hasn't read that research recently, so it needs help and it needs a reminder. It can design a step-by-step financial playbook. If you give it a complete picture of your situation, using the prompts we covered today, it can build you a prioritized, logical sequence of financial actions. Pay off this debt first, max this account first, build this reserve first, the logic is sound, the sequencing is often correct, and the output is something you can actually act on. A lot of people have never had that all in one place before, and it's a special moment. And this might be the sleeper hit. It's available at 2 a.m., when you're anxious and your advisor is asleep and the urge to actually do something financially catastrophic is at its peak. The best use of AI may simply be, have the conversation before you make the move. Let it ask you the question, what would have to be true for this to be a good decision and then sit with the answer. But here's the thing I want you to actually take away from this episode, the thing that hopefully ties all of this together. AI is a tool. It is an extraordinary tool, and I think it is capable of exceptional things. You need to learn it today, but the tool is only as good as we are at using it. Thus why I felt the need to design this episode for all of us. Garbage in, garbage out, and in financial planning, garbage can be measured in the size of your retirement account. So use the prompts I gave you. Be honest with your inputs. Verify the specific numbers, understand the privacy trade-offs, and treat AI as a brilliant, well-read, occasionally overconfident friend who has read every book on personal finance, but not as a licensed advisor with a fiduciary duty to you who is up to date on all current data. The best financial decisions I've seen people make are the ones where they understood why they were making them. AI can help you get there. The understanding still has to be yours. Alright, that's all I've got for this week. I hope it was somewhat useful. And if it was, you know what I'm going to say. Apple, Spotify, Review, 45 seconds, and it genuinely means the world to me and helps us get these notes out to an ever greater audience. As always, I hope this gives you something useful to think about in the week ahead. Thanks for tuning in to your money guide on the side. If you enjoyed today's episode, be sure to visit my website at TylerGardiner.com for even more helpful resources and insights. And if you're interested in receiving some quick and actionable guidance each week, don't forget to sign up for my weekly newsletter where each Sunday, I share three actionable financial ideas to help you take control of your money and investments. You can find the sign up link on my website, TylerGardiner.com or on any of my socials at social cap official. Until next time, I'm Tyler Gardiner, your money guide on the side. And I truly hope this episode got you one step closer to where you need to be.