The Experiment Begins
I’ve always been one of those people who says they’re “pretty good with money.” You know the type. I have a checking account, I pay my bills on time, I don’t carry a balance on my credit card. At least not most months. I figured I was doing fine, honestly. So when YNAB rolled out their new Reflection feature in the fall of 2025, which uses AI to automatically categorize your spending and generate these weirdly personal monthly summaries about your financial personality, I thought it would be amusing. A fun little tech experiment. Something to maybe write about.
I wasn’t prepared for what it would actually show me.
The setup was straightforward. I linked my accounts, let the AI do its thing, and committed to tracking every single transaction for 60 days. And I mean everything. The three-dollar coffee. The impulse bookstore visit. The “I’m tired and don’t want to cook” delivery order at ten PM. All of it. I thought I knew what I’d find. Spoiler: I didn’t.
What the Numbers Actually Said
Two months of data doesn’t sound like much until you see it laid out in front of you. The AI didn’t judge, which somehow made it worse. It just categorized and summarized with this unsettling neutrality. One category in particular stopped me cold: discretionary spending. The stuff that wasn’t rent, wasn’t utilities, wasn’t truly necessary. I’d estimated maybe fifteen percent of my monthly income went there.
The actual number was thirty-one percent.
Thirty-one percent. That’s nearly a third of my money going toward things I didn’t plan for, didn’t remember half of, and couldn’t quite justify when I saw them all lined up. The delivery orders alone were shocking. When they’re individual transactions spread across two months, they feel manageable. When they’re aggregated into one line item labeled “Convenience Food,” they look less like occasional treats and more like a habit I’d been ignoring. The coffee runs had their own moment of reckoning too. I drink coffee most days. Most expensive days, apparently.
What got to me wasn’t the shame, though there was plenty of that initially. It was the recognition that I’d been lying to myself in the most mundane way possible. Not dramatically, not criminally, just casually. The way you might tell yourself you’re eating healthy while eating salad with a pint of ice cream for dessert. The lie felt smaller than it was.
The Humiliation Was Educational
Here’s what I wish someone had told me: you have to let yourself feel bad about this stuff for a minute. People online will tell you not to shame yourself about spending, and they’re right in theory. But there’s a difference between shame and clarity. Shame is useless and sticky. Clarity just stings at first.
I looked at the data and had to confront the fact that I’d been operating with a story about myself that wasn’t true. I’d built an identity as someone financially responsible, and then I’d spent thirty-one percent of my income on discretionary stuff without ever acknowledging the pattern. The AI didn’t do anything miraculous. It just made the invisible visible. It’s surprisingly humbling to realize your own brain has been doing math you wouldn’t approve of.
The Financial Health Network reported in 2025 that forty-three percent of Americans describe themselves as “financially vulnerable,” the highest number they’ve recorded since they started measuring. That statistic kept running through my head while I stared at my own numbers. I’m not in that category by their definition, but I was vulnerable to my own self-deception, which felt like a different kind of reckoning.
What surprised me was that the hard part wasn’t seeing the numbers. The hard part was not immediately swerving into some extreme correction mode. My instinct was to swing all the way to the other side, to cut discretionary spending to nothing, to punish myself via austerity. That impulse felt familiar and wrong. I had to sit with it.
The Data Changed How I Actually Behave
Here’s where the story becomes less about humiliation and more about what happened next. After I spent a week feeling appropriately horrified, I started reading YNAB official feature announcements and user data about how other people had handled similar discoveries. Turns out there’s research on this. A 2025 study from Intuit found that sixty-five percent of Americans who stick with budgeting apps for ninety days reduce their discretionary spending by an average of nineteen percent. That’s not punishment territory. That’s reasonable adjustment.
It turns out that new YNAB users save an average of six hundred dollars in their first two months and then six thousand dollars over their first year when they actually engage with the tracking process. Those aren’t people who cut everything. They’re people who got honest about what they were doing and then made small, sustainable changes.
I’m only two months in, so I’m not hitting those savings numbers yet. But I did something that felt more important: I reset my relationship with the discretionary spending category. Instead of pretending it didn’t exist or making it my enemy, I acknowledged it. I moved it from thirty-one percent to twenty-four percent over the next month, not by depriving myself, but by being actually conscious about choices. The difference between choosing a coffee and autopiloting to the cafe is bigger than it sounds. Same with delivery. Sometimes it’s the right call. I just needed to know I was making it rather than discovering it in my bank statement three weeks later.
Why I’m Telling You All of This
I’m showing you my numbers because I think a lot of us are walking around with the same quiet delusion. We’re doing better than average, we’re not drowning in debt, we’re managing. And maybe we are. But maybe we’re also just not looking too closely. The average American household was carrying ten thousand, four hundred seventy-nine dollars in credit card debt as of late 2025. That’s real. That’s a thing happening. But the thing that gets me isn’t the statistics. It’s the possibility that you, reading this, might have your own version of this story hiding in your own numbers.
The Reflection AI feature was supposed to be a convenience tool. Automatic categorization, personality summaries, that sort of thing. What it actually became for me was a mirror. Not a harsh one, just an accurate one. It showed me what I was doing without the filter of how I preferred to see myself.
You don’t need an AI feature to do this for yourself. You just need honesty and maybe sixty days of paying attention. The Financial Health Network 2025 U.S. Financial Health Pulse shows that most of us are in some state of financial vulnerability, regardless of income level. We’re all working with incomplete information about our own behavior. All we’re really doing here is trying to complete the picture.
So I’m curious. What number would you find if you actually tracked everything? What’s the category that would surprise you? I’m not asking to shame you. I’m asking because I’m pretty sure you’d be in good company, and sometimes that makes the reckoning feel less lonely. Let me know what you find.