You Asked ChatGPT About Your Money. Good.
Somewhere in Columbus this Friday night, someone will be lying awake over a money question that has followed them around all week. And instead of letting it spin, they will pick up their phone and ask ChatGPT or Claude, and go back and forth with it for an hour, and fall asleep knowing more about their own RSUs than they did at dinner. I want to say something my industry mostly will not: I think that is great. Genuinely. This issue is about what to do with that conversation the next morning.
WHY I LIKE IT WHEN YOU DO THIS
AI is awesome at what it is awesome at. It takes in a question and hands back more organized information than any human could assemble at midnight, and most of it is accurate or at least directionally accurate. When clients research this way, they show up to meetings with stronger, sharper questions, and the whole conversation starts three steps ahead. My only ask is the same one I would make of any homework: send me the thread before we meet. Context matters, and meeting preparation is still part of my job. A client who arrives curious and pre-read is not a threat to what I do. They are the best version of the meeting.
THE TWO KINDS OF KNOWLEDGE
At a conference this month, a presentation sent me down a rabbit hole that traces back to an economist named Friedrich Hayek, who wrote in 1945 that the world runs on two different kinds of knowledge. There is general knowledge, the textbook kind, rules and formulas that are true everywhere and can be written down. And there is knowledge of particular circumstances, the local kind, what is true about this person, this family, this moment, which never makes it into a textbook because it lives in exactly one place. Swap in modern words and you get the split that matters right now: there is textbook knowledge, and there is judgment. Both are real knowledge. Only one of them can be scraped, trained on, and served back to you at midnight.
WHERE THE RULEBOOK ENDS
AI has read every rulebook ever written, and it will read them to you beautifully. It will tell you exactly how RSUs get taxed. It will not tell you your tax bill, because your tax bill lives in your particulars: the vest schedule, the bonus, your spouse’s income, the withholding gap quietly building since July. It will tell you that a 4 percent withdrawal rate is considered safe, and that is a fine reading of the textbook. What it cannot know is that the person asking is 70 with declining health, and that for their actual life expectancy, 4 percent might be far too conservative, that the "safe" answer is the one costing them the trips and the giving they would trade almost anything for later. Notice the direction of that mistake. Everyone assumes the human’s job is to talk you out of the machine’s risky ideas. Just as often it is the opposite. The rulebook can read you the rules. It has no chapter on you.
WHO DO YOU GO TO FOR JUDGMENT
Here is the question the AI era actually raises, and it is not "should I use these tools." You will, and you should. The question is: when the answers are everywhere and half of them are slop, who do you go to for judgment? Friends? Family? An advisor? The honest answer is that everyone needs a network of trustworthy people, built on purpose, before the decision that needs them. That is what I mean when I say I want to be a thinking partner. Bring me the machine’s answer and I will tell you where it is right, where it is incomplete, and what it could not have known about you. And for transparency, AI is remaking my side of the desk too. Over the next few years it will absorb the data entry, the paperwork, the scheduling, the monotonous middle of this profession, and I am glad to hand it over, because every hour it takes is an hour I get to spend actually present in a meeting. If a client trusts your judgment, AI is a tool, not a threat. That sentence is where this whole profession is headed.
A SIMPLE EXAMPLE
Maya is 33, a product manager in Columbus with RSUs vesting every quarter. Before our meeting she spent a weekend going back and forth with an AI about whether to sell her shares at vest, and she sent me the whole thread, which made my week. The thread was good. Sell at vest logic, correct. Taxed at vest not at sale, correct. Diversification reasoning, genuinely well put. Then we added the chapter it could not write. Her real bracket sits at 35 percent against the flat 22 percent her vests withhold, which means the projection, not the rulebook, says an estimated payment lands in January. She wants a house in Clintonville inside two years, which turns her vested shares into a down payment timeline rather than a philosophical debate. And she said out loud something no chatbot elicits: she would never forgive herself for betting the house fund on her employer’s stock, however much she believes in the company. The AI got her 80 percent of the way there, and I told her so. The last 20 percent was not information. It was her.
THE WHOLE IDEA
Keep asking the machine. Ask it everything, at midnight, in the car, mid meeting if you want. Then bring the answer to someone whose whole job is the last mile: your tax bill instead of the tax rules, your one life instead of the average life, the amendments real life writes into every textbook answer. The rulebook is free now, and that is a gift to everyone who is curious. Judgment is the service. Choose where you get yours on purpose.
CLOSER TO RETIREMENT? FILE THIS AWAY
The two kinds of knowledge age in an interesting way: the closer you get to retirement, the more your questions become the particular kind. Whether to take the pension as a lump sum, when to claim Social Security, whether 4 percent is too much or, just as often, too little for your actual health and your actual goals. AI can read you every rule of thumb ever published, and rules of thumb are averages, and nobody retires as an average. So use the tools freely, and at the same time build the human side of your bench: the advisor, the CPA, the estate attorney, the friend who tells you the truth. Assemble them before the year you need them, and let every rule of thumb be a starting point instead of a verdict.
WORTH A READ
The Use of Knowledge in Society, F.A. Hayek (Econlib)
Artificial Intelligence and Investment Fraud: Investor Alert (Investor.gov)
Tax Withholding Estimator (IRS.gov)
If there is an AI thread about your money sitting on your phone right now, send it over and let’s finish it together. That is not a gimmick. It is honestly my favorite kind of meeting prep.
Curious what working together looks like? There is more at jcsretirementtaxadvisors.com, and if a conversation sounds easier, you can grab 30 minutes on my calendar whenever it suits you. No pressure, just a conversation.
Know someone who should be reading this? Forward it along, that is how most people find me. And if this was forwarded to you, one subscription at buckeyeretirementbrief.beehiiv.com gets you both briefs: retirement on Mondays, equity comp on Thursdays.
Jesse Stacy MTAX, CFP®, EA
JCS Retirement Tax Advisors
This newsletter is for education only and is not tax, legal, or investment advice for your situation. The example is illustrative. AI outputs about tax and investments can be incomplete or wrong for your circumstances; verify anything that matters against your own numbers or with a professional before acting.