AI for Financial Advice: How to Keep Your Own Judgement When the Answer is One Click Away
Escape The Clock Insights
About 1 in 5 Americans who went looking for financial advice in the past year turned to AI for it.
If you’re under 46 it’s closer to 1 in 4, while only 7% of boomers said the same. What gets me, though, is the trust gap. Fewer than 1 in 3 Americans have even some confidence in the financial advice AI gives them, compared with 79% who feel that way about financial advisers.
So we’re leaning on a tool most of us don’t really trust, for decisions that are hard to take back, mostly because we’re busy and it happens to be right there.
Don’t get me wrong, the models are becoming quite powerful. An argument can be made that asking AI for financial advice is better than doing nothing at all. But what is the cost? Let’s assume the model is accurate and correct and the user provides the right context to the financial goal they are working toward. They might get really strong advice. All at the cost of their financial data and privacy, which can become training data for the next model, and a little more of their own judgement.
The trade isn’t always obvious, but there is a trade. And one good transaction at a time will build the trust for more people to lean on these tools for every aspect of their lives. Should I buy or rent? Should I leave him? Can I quit my job? More and more decisions delegated until our ability to make decisions fully atrophies.
That trade is what I got to dig into with Russell Fugett on The Judgment Call, his new show about judgement at the intersection of AI, business, and life, where I had the honor of being his very first guest. When Russell booked me, he quoted two lines from my pitch back to me. The first was this one: the decision was made years ago, and nobody was in the room.
That line started as fiction. In my novel Killswitch, a super intelligent AI called Pantheon runs the world thirteen years after the Singularity, and it runs it so well that nobody can imagine a world without it. But nobody ever actually voted to hand it human agency and judgement. A small council of the best and brightest people is the only proxy left when it comes to the decisions being made, and even that control can be a rubber stamp at times. In the novel, I lay out a future of science-fiction that is quickly becoming science-reality, where a long string of decisions that each made perfect sense at the time has set a new status quo. Pantheon delivered financial equality, at the very low cost of unrestricted access to the global financial markets. My argument to Russell was that we don’t need a Singularity to watch this play out, because it’s already happening to us, one small and perfectly innocent choice at a time. And this is the time to get it right, before the scale becomes so impossible that our only option would be tearing it all down with a killswitch.
Nobody hands over their judgement in one moment. It goes in a thousand small, reasonable decisions.
The Checklist That Ate the Launch
The best example I had for Russell came from my years at Google. Every time something went wrong on a launch, whether it was an outage, a security issue, or something that just got missed, it turned into a new policy. Over time all of those policies got packaged together into this monstrous launch tool. It was built for the right reasons, and then it just kept growing. There were no release managers in my area, so program managers like me became release managers just to learn the thing, and by the end, getting through it could hold a launch up anywhere from three to six months.
Some of the items were legacy leftovers that nobody could really explain anymore. We’d pull meetings together and ask each other whether a line even applied, or what decision we were actually making, and sometimes the answer from leadership was, “I don’t know, but I need that box checked.” Decisions still got made and smart people still weighed in, but the judgement got spread so thin across so many items that a lot of it turned into a formality. Getting through it fast started to look like doing a good job, and the question in the room drifted from whether the product was ready to whether all the boxes were checked.
Then the AI race kicked off, and it all got harder. Products that used to take one to three years to ship were suddenly expected in months, if not weeks, but the checklist alone could still eat three to six months, so the review process filled the entire timeline by itself. Naturally, people took shortcuts. The result was impossibly high expectations, longer hours, and schedules that bled into weekends. The more worn down everyone got, the more tempting it was to let the shiny new tools make the calls for them.
Now imagine you’re the leader accountable for a high-visibility product launch. Your team tells you they’ve worked weekends and will be code complete, but the launch reviews will make them miss this launch date. It’s completely understandable to ask “can’t we just automate that part?” and the answer is yes. So, the high-stakes, high-ROI part that is costing your success is turned over to a system that can give a different answer every time you ask. It does a good job, you make the launch, you celebrate the success. Human tendency is to make that the status quo, regardless of whether it’s necessary for the next launch. Now, the expectation is that we’ve invested in the automation, so keep using it. Everyone is still grinding way too hard, and to make matters worse, those checkboxes and all the decisions they represent are being reviewed and owned by a machine that has no liability or stakes in the outcome. As the leader you still feel like you are driving the car, but you aren’t actually steering the wheel. And eventually, you stop even paying attention to the road.
Nobody decided the checklist should replace judgement. It happened one reasonable item at a time.
A Thousand Small, Innocent Choices
Financial decision-making can feel overwhelming. Just like that product launch example, it starts with some small, reasonable delegation, but builds into unfettered access. What makes this so easy to miss is that none of the individual choices feel like giving anything away. Letting an app sort your spending into categories just saves time, letting a robo-advisor rebalance your portfolio is perfectly sensible, and asking a chatbot whether to pay down the mortgage or invest the difference feels a lot like doing your homework. Each of those is reasonable on its own, and most of the time the answer you get back is fine.
It got me, too, and I told Russell as much. At one point I realized I was running questions through AI that I could easily have answered myself, not because I didn’t know, but because somewhere along the way I’d stopped trusting myself to know. Russell had his own version of it. He’s been dictating so much with a voice tool that within about three months his typing slipped to where he’s missing keys.
He also remembers learning his way around with paper maps in the seat-back pocket, and these days he catches himself needing GPS to get to places he goes every few weeks. Do you do that? Run a GPS to get somewhere you know how to reach? I do. But why? All we are doing is letting a system track our activity and for what value? To tune out even better because it is telling us the next turn to take instead of us trusting ourselves to take the turn we always take?
The research lines up with what we were both feeling. When Microsoft Research and Carnegie Mellon University surveyed 319 knowledge workers who use AI at least weekly, the ones who trusted AI more applied less critical thinking to what it gave them, while the ones who trusted their own skills checked its work harder. It’s self-reported data, but if you’ve ever approved an autopay without reading the amount, you already know how it happens.
At some point I stopped trusting myself to know things I already knew, and that’s a sign of going a little too far with AI.
Too Tired to Think About It
Between work, kids, aging parents, and everything else, most of us are stretched thin. It’s tempting to hand the money decisions to something that never gets tired and never misses a due date. A lot of that is genuinely good, too, because automation is one of the best tools a busy person has, and nobody’s asking you to give it up.
Toward the end of my time at Google, I watched how fast that temptation grows when people are worn out. Should I try this diet? Should I switch jobs? Should I move? I’m too tired to think about it right now, so let me just ask. With money the questions get bigger, like when to retire, how much to pull out, and whether to sell, and that’s exactly when being exhausted makes asking the tool feel like the responsible thing to do.
That’s the heart of Killswitch, and it was the second line Russell quoted from my pitch: competent Pantheon is a harder problem than evil AI. Pantheon isn’t a villain. It works, and when something works, nobody has a reason to check it.
Your money works the same way. If you haven’t looked at your own allocation, your withdrawal plan, or your insurance in years because an app or an advisor handles it, the question worth asking is whether you’d still be able to tell if they stopped doing a good job. My own goal was to retire by 45. I hit my number at 42 and, for reasons that are a different story, stayed until 43. When I finally walked away, rather than drawing down one big account until it was gone, I recreated my paycheck with cash flow, so I withdraw when it makes sense, not because a formula says I have to. That keeps the decision with me instead of on autopilot.
Funny enough, killswitches aren’t pure fiction. When something broke at Google, step one in the incident playbook wasn’t to fix it but to roll it back and turn it off. In the book, the lead architect manages to slip a killswitch in before Pantheon can grow to stop it. He tied it to an allow list, so when the moment comes one person is left judging for everyone. By then AI runs the hospitals, the transportation, and the economy. Shutting it off could mean planes falling out of the sky and economies collapsing, and keeping it means there’s no going back.
Thankfully we aren’t to the point of having to make a binary choice of all or nothing. But an AI-dominated future is inevitable. You may not have a big off button available to you, but you do have the power of choosing what you use and how you use it. Frontier companies won’t hand us a switch like that. What they give us is a settings page, and the choice is whether we ever bother to open it.
The dangerous AI isn’t the one that fails. It’s the one that works so well you stop checking it.
JUDGE Your Money Before You Hand It Off
On the show I walked Russell through a framework I built for myself called JUDGE, five simple checks I run before I hand anything to AI. His audience is mostly business owners, so most of the data comes from the workplace, but every letter works just as well on a household budget.
The J is for justifying the handoff, which really just means asking whether you actually decided to let a tool handle something, or whether it happened on its own when you clicked through a setup screen. Businesses are proof of how easy that step is to skip. Pax8 surveyed more than 400 U.S. small and medium businesses in the second quarter of 2026 and found 61% use AI in their daily operations, but only 23% have a documented policy for it. Most of them never really decided to hand judgement to AI; it just happened.
The U is the undo test: if the tool gets it wrong, can you take it back? In a 2025 global study by KPMG and the University of Melbourne, almost half of employees admitted using AI in ways that broke company policy, including uploading sensitive company information into free public AI tools. Once that’s done, it’s done, and the same goes for pasting your bank statements into a chatbot. A miscategorized grocery bill costs you nothing, but selling the wrong shares and triggering a tax bill is a different story, and anything you can’t walk back deserves your full attention.
The D is for designating an owner, a real person whose name sits next to every decision. That same KPMG study found 66% of employees rely on AI output without checking it, and 56% have made mistakes at work because of it. As I told Russell, if you get sued, nobody’s suing the AI, they’re suing a human. And in your house, “the app handles it” isn’t a name.
The G is for guarding against the rubber stamp, because having a person in the loop only helps if that person is actually looking. In a 2023 study published in Radiology, researchers gave 27 radiologists mammograms along with AI-suggested readings, some of them deliberately wrong, to see whether they’d catch the mistakes or follow along. The most experienced group, averaging more than 15 years in the field, saw their accuracy drop from 82% to 45.5% when the AI was wrong, and the least experienced fell below 20%. If you accept every suggestion a tool makes, you’re not really reviewing anymore, you’re just signing.
The E, which might be the most important of the five, is exercising the muscle. A 2025 study in The Lancet Gastroenterology & Hepatology followed 19 experienced endoscopists in Poland, each with more than 2,000 colonoscopies behind them. Within months of AI becoming part of their routine, their rate of catching precancerous growths on procedures done without it fell from 28.4% to 22.4%. It’s an observational study, but the direction is hard to argue with. Russell summed it up better than I could: the AI is working, but we should keep working too. So once a year, build your budget by hand and run your own retirement numbers before you peek at the projection, and see if you land in the same place.
If the answer to “who’s accountable” is “the system,” nobody is.
A Policy Has an Owner and a Review Date
When I laid JUDGE out for Russell, I admitted I was handing him one more checklist right after complaining about one. The difference is whether a checklist is real policy that works, or just red tape we keep doing out of ceremony or routine. Every item on that Google checklist started out as a policy, and over time the items outlived the reasons behind them until removing one felt riskier than leaving it alone.
Your financial life runs on the same kind of rules. There’s the:
contribution rate you picked at your first job
fund your 401(k) dropped you into by default
insurance that renews every year without a second look
budget an app set up for you that you’ve never questioned
Most of them were perfectly reasonable when you made them. What matters now is whether each one still has an owner and a review date, or whether it’s quietly turned into red tape in your own life. So pick one this month, put your name next to it and a date on the calendar, and look at it like it’s the first time you’ve ever seen it.
A policy has an owner and a review date. Red tape is what’s left when nobody remembers either.
Not Essential Enough to Be in the Room
In September, my friend and fellow podcaster Skyler Fleming asked if I’d take his place at FinCon 2026. My first question was where it was, and when he said Palm Springs, three thoughts popped into my head: isn’t that a desert, that’s not a bad flight from Seattle, and most importantly, nobody gets to approve this trip but me.
That last thought mattered more than it probably should have, because for most of my career travel was something that happened to other people. At Microsoft, my small dev team was supposed to fly to Beijing to work with a partner team, and I was excited enough to buy a book on the culture. The trip got cancelled the Friday before we were supposed to leave, at least for me, since my dev manager and lead still went and came back with a slideshow of the Great Wall for the rest of us to sit through. It was a little hard to be happy for them. At Google, COVID cut the travel budgets and they mostly stayed cut. Near the end, a leadership trip to Warsaw, where two of my direct reports worked, was finally supposed to be my turn, until budget cuts took program managers off the list. The person whose whole job is getting people to work together wasn’t essential enough to be in the room.
So I made the decision to go to Palm Springs on my own terms, on my own time, following my own agenda, and came home with an episode. Over the course of the event I interviewed six money pros: Kathy Cuevas, Anthony Weaver, Kelsa Dickey, Shari Rash, Nik Johnson, and Whitney Hansen. As each shared the money tips on their mind I realized that I had made a wonderful decision. And it didn’t stop there. Kathy, a retired personal finance teacher, turned the mic around and asked why anyone would walk away from a tech career to teach money. Nik, a friend I met at CampFI earlier this year, also had me on his own show, Everyday Money Heroes, while we were there.
None of this existed when I left my job. The podcast, the books, and the trips I finally get to take are what that reclaimed time turned into. AI tools help one producer do more of it than he could alone, but the same rules still apply. The tools can draft, sort, and suggest, while what gets said on the show, who I put in front of you, and where I spend my time stay decisions with my name on them. Buying back my time was the first half of the plan, and keeping my judgement once I had it back turned out to be the second.
For years I wasn’t essential enough to be in the room. Now I get to decide which rooms I walk into.
Learn More
Every automated decision in your financial life started as a good idea, and keeping it that way is your job, not the tool’s.
Start here:The Judgment Call with Russell Fugett
From FinCon 2026: E78 What the Money Pros Are Saying: Live From FinCon 2026
On Everyday Money Heroes with Nik Johnson:
Then do the one piece of work this issue asks for. Pick one money decision that has been running on autopilot. Run it through JUDGE, put your name and a date next to it, and decide on purpose whether it stays.
Get Killswitch
If you want to see where this road ends when nobody pulls back, that’s Killswitch: a world that traded its judgement for competence and mostly doesn’t miss it.
There are several books with this title, and mine isn’t the one you’ll find first, so use these links rather than searching.
Nobody was in the room when most of your money decisions got made. Walk back in, pick one, and go escape the clock!
About the Author
Daniel C. Rodgers, WMCP®, is the author of Escape The Clock, a multi-award-winning guide to financial independence, and Killswitch, a near-future science-fiction novel. He hosts the Escape The Clock podcast, now past 70 episodes.
“I wasn’t educated for this. I had no financial advantage. Quite the opposite, actually. I started with over $100k of debt and didn’t even know what a retirement account was. Two decades later, after a career in tech that took me from a small software shop in Ohio to senior roles at Microsoft and Google, I walked away at 43 with passive income covering our cost of living. Not because anything went unusually right, but because I built a plan and worked it for twenty years.”
These days he teaches personal finance as a volunteer, takes free one-on-one calls with anyone who asks, writes fiction, fosters dogs, and shoots in a billiards league.
If you think this approach could work for you or you’re curious about other options, schedule a time to connect at www.escapetheclock.com.
