Can AI Be Your Financial Advisor?

A few years ago, if you wanted financial advice, you had a fairly predictable set of options. You could call your financial advisor, of course. You could call your dad. You could ask the guy at work who owns three rental properties and somehow manages to bring them up during every conversation. Or you could spend Saturday night wandering through the financial wilderness of the internet where eventually someone would explain why the global economy was about to collapse and you need to buy gold and hedge and... gosh. Ugh, gosh. So there is another option. a lot more popular right now. You can ask a machine. You know, before it was, I maybe just googled some topics and found some websites, but now some of that information gets consolidated into artificial intelligence.

And this is what people are doing right now. In a 2026 Gallup poll, it found that roughly one in five Americans who sought financial guidance in the previous year used AI as one of their sources. Younger Americans were substantially more likely to use it. Despite that, only 3% of adults said they have a great deal of confidence in AI's ability to provide financial advice. So we've arrived at a strange moment. People are asking AI what to do with their money while simultaneously admitting they don't really trust the answer. As people become more comfortable using and relying on AI, I expect those numbers to increase, even if the value of the advice doesn't.

You're listening to the Thoughtful Investor podcast, brought to you by Yach Advisors. I find all of this fascinating because I spent more than 15 years working as a financial advisor, and over the last few years I've started using artificial intelligence. I'm probably more optimistic about this technology than some people in my profession. AI can read thousands of pages faster than I can read one. It can explain Roth conversions ten different ways without getting annoyed. It doesn't get tired. It doesn't have a 4.30 meeting that I have to attend. And it'll probably happily discuss the finer points of modern portfolio theory at two in the morning without getting annoyed. I can't really say that for most people that I know.

So I wanted to ask the question that's admittedly a little uncomfortable for someone in my profession. Can AI actually be your financial advisor? Coming from a financial advisor, you might not expect my answer. Bear with me here. So researchers have already started putting together this question into different testing scenarios. In one study, researchers asked 32 large language models to construct portfolios for 64 different investor profiles. They changed things like risk tolerance, investment experience, sustainability preferences, gender, country of residence. and the machines did pretty well. The models generally recognized meaningful differences between investors and adjusted their recommendations accordingly. The historical performance of the resulting portfolios was even comparable to professionally managed benchmarks.

That should make a lot of people in my profession uncomfortable. because of financial advice Because if financial advice simply means answering questions, I'm 45 years old, I want to retire at 65 for example, and I have $500,000, what should my portfolio look like? Then yeah, AI is going to become pretty good at answering that question. And it's going to cost a lot less money for them to build a model portfolio for you. And I don't think advisors should dismiss that and try to just point at the things AI gets wrong. Technology doesn't have to replace everything an advisor does to radically change the economics of financial advice. If software can perform 70 or 80% of the analytical work that was once required by a trained professional, that's enormously consequential.

Some services people have historically paid thousands of dollars for may eventually become nearly free. I mean, that's that's what happens as the profession evolves. I mean, that's what happens as the profession evolves. We don't need to bring out a 500 page document of prospectuses and comb through them in order to provide advice for clients. We can cut that out because of computers. It made it a lot easier and a lot more efficient. And who benefits in the end? Well, costs came down considerably in the financial services industry. So at the end of the day, it was the consumer that benefited the most. But after spending most of my career sitting across the table from families, I've come to believe that the hardest question is helping you understand what your deepest questions are even before you do.

I'll explain. A client walks into an office and says, I want to retire at 62. A client walks into an office and says, I want to retire at 62. That's the prompt." except maybe it isn't. He has a folder under his arm, a few statements, a yellow legal pad with some numbers scribbled on it. He spent 20 years thinking about retirement as a date on a calendar. Then you start talking. His wife wants to retire at 58. Maybe they didn't have that conversation. They have a daughter who's struggling financially. His mother may eventually need long-term care. Most of his net worth is tied up in the company he works for.

He's terrified of another 2008. He lived through it. He watched his father panic and sell everything near the bottom, never recovered. And he's comfortable with the risk, but when his portfolio falls 12%, he loses sleep at night. He's lost his nerve in investing. And somewhere in that conversation, retirement at 62 stops being really the question he came in with. That's financial planning. The numbers matter. but eventually the spreadsheet runs into a human being. Human beings have emotions. Human beings want security. Researchers are starting to find this boundary too. A 2025 experiment involving 64 participants found that AI advisors could sometimes reform about as well as humans that about as well as humans at eliciting investor preferences, but the models struggled dramatically when people's needs conflicted.

More troubling, participants could develop greater trust and satisfaction with more extroverted A.I. personalities, even AI personalities, even when the models were giving worse advice. So people's satisfaction with their advisor had less to do with the advice and more to do with the feeling that they came out with. And as an advisor, that's scary to me because being confidently wrong is one of the worst thing an advisor can do. I think everyone's had that experience. They walk into customer service and they get wrong advice. They know it's wrong, but the person behind the desk is so confident about it, you start to second guess yourself. Even if you know the answer, think about that for a moment.

The machine doesn't necessarily have to be right. It just has to sound right. And that's not entirely an artificial intelligence problem. It's a problem that we have that's been around forever. We like confidence. We like certainty. We're always drawn to people that tell us they know what's going to happen. Wall Street figured this out long before Silicon Valley did. AI just simply made confidence scalable. There's another problem that anyone who uses these systems eventually will encounter. The dangerous thing about bad financial information isn't that it's obviously wrong. It's that most convincing mistakes often contain just enough truth to sound right. AI can give you a wrong answer in a beautifully written six paragraph explanation, complete with bullet points, citations, reassuring conclusions.

And unless you already know enough to question it, you probably won't notice. I'm an expert on financial planning and investments. But if I ask a question about how to rebuild an engine, I would But if I ask a question about how to rebuild an engine, I would really hope that they have everything buttoned up there or else the consequences could be drastic. There's also a difference that's less philosophical and more practical. accountability. If AI gives you bad advice, who's going to hold it responsible? Financial professional operates within a regulatory framework, Documents recommendations has a professional obligation surrounding the advice given. hopefully like a certified financial planner, a fiduciary responsibility to put the interest of the clients ahead of their own.

None of that guarantees good advice, of course. Human beings are capable of giving bad financial advice, but accountability still matters in the decisions involving someone's life savings. I don't think the lesson here is don't use AI for finance. In fact, I would probably say the opposite. I think AI may eventually democratize financial knowledge in ways that would have been almost unimaginable when I started in this business. Someone who could never afford a financial planner can ask, "What's an expense ratio? How do I avoid a wash sale?" A young employee can make sense of alphabet soup inside a 401(k). A family can financially understand the difference between a traditional and a Roth IRA without spending an afternoon lost on Google.

That's not a threat to financial planning. That's progress. A world in which people are more informed is a better world. AI can help professionals as well. Let the machines analyze thousands of possibilities. Let it organize information. Let it research tax law. Let it run scenarios, challenge assumptions, handle some of the work machines are exceptionally good at. Then they sit across from a human being and do something a little old-fashioned. They listen, ask the questions they hadn't thought to ask. Notice when the husband says he's comfortable with the risk and the wife quietly looks down at the table. Understand, the house isn't simply an asset worth $850,000. worth eight hundred and fifty thousand dollars.

It's a place they raise their children. There's meaning there. Recognize that mathematically optimal and emotionally survivable things are two different portfolios. Knowing the client asking, do I have enough? isn't really asking for a number. They're looking for reassurance. And when the market falls thirty percent and every headline in When the market falls 30%, every headline in the world is telling your client, "This time is really different. This is the one." As an advisor, you can remind them about the conversation you had five years earlier when the world felt considerably safer. Will an A.I. be able to remember that conversation? Will an AI be able to remember that conversation? Will an A.I. be able to remember that conversation?

Having interacted with AI myself, I don't think it remembers a conversation I had 10 a conversation I had ten minutes ago. That's difficult to put into an algorithm, an entire history and an entire understanding between two people. Maybe someday it'll become better at that. I wouldn't bet against technology. But here's where I've landed for now. If you want to understand an investment concept, AI can be a great teacher. If you want to brainstorm financial possibilities, it can be a great research assistant. If you want to organize your thinking before meeting with an advisor, I think that could make the meeting considerably better. Putting in a prompt that says, "I don't know what to talk about.

Give me 10 questions," probably isn't very productive either because it might not be 10 questions that you really need to know or want to know. It just creates extra work and distracts from the things that are most important. A good advisor with no context will be able to draw some of the drivers out of you. Drivers, meaning things that are really things that you're deep down concerned about. I'll give you an example. Someone asks, "I want to make sure I can retire at 62." What they're really asking is, "I don't want to run out of money like my father did. If I live to 95, I don't want to be a burden on my children." Those are really the feelings behind the question.

AI is not going to draw that out of you. To use a bit of a cliche here, the outputs are only as good as the inputs. But if you're asking a machine to make consequential decisions about your retirement, your family, your taxes, your estate, or your life, remember what you're actually talking to. It doesn't know you. It knows what it's been told. Those aren't necessarily the same thing. LLMs large language models are designed to LLMs, large language models, are designed to just complete thoughts one word at a time, and sometimes it can ramble into nonsense. We see it all the time. For centuries, technology has made humans' lives easier. They replaced muscle.

Calculators replaced arithmetic. Computers replaced filing cabinets. And we've had to reconsider what our time is actually worth. One of the first inventions that we think about in the history of mankind is fire. What did fire do? It cooked food so that we didn't have to spend hours and hours chewing raw meat. We could cook it and eat it a lot faster. What did we do with that extra time? We looked to the skies. We made tools. We thought we developed philosophy, language, science, writing, all of that. It's fascinating what giving us extra time does to the development of mankind. Perhaps the future of financial advice won't be less human because of artificial intelligence.

Perhaps it'll be more human. Maybe we'll have time to develop our social skills, to get back to the root of what makes us human. And I think that might be good for everyone. I'm Bryan Yach. This is The Thoughtful Investor. I'm a wealth advisor with Yach Advisors and a certified financial planner professional. Follow us for more investing topics and visit our website, yachadvisors.com. That's Y-A-C-H, advisors.com. The Thoughtful Investor is brought to you by Yach Advisors. Copyright 2026, all rights reserved. Yach Advisors' registered branch office is located at 2241 East Continental Boulevard Suite 130, Southlake, Texas, 76092. Security is offered through Cetera Wealth Services, LLC. Remember FINRA, S-I-P-C. Advisory services offered through Cetera Investment Advisers, LLC, a registered investment advisor.

Cetera is under separate ownership from any other named entity.