“Trust Me” Versus the Data
A bike-sizing problem that keeps showing up in my work.
Two confident answers, opposite directions. The salesperson said trust me. The machine showed its work. I keep meeting this exact problem in my actual job.
I’m buying a bike. I found the one I want, and now I’m stuck on a question that sounds small but isn’t… large or extra large.
So I did what most of us do now. I asked the dealer, and I asked AI. The dealer rep sells these bikes every single day. He’s moved thousands of them. He is 100,000% certain I want the extra large. AI looked at my height, my inseam, the geometry of both frames, and a few of my own quirks, and it landed firmly on the large. Two confident answers, pointing in opposite directions.
Here’s the part that’s been bugging me. The dealer’s whole case was “trust me, I’ve done this all my life.” Which is fair. He has. But so have I. I was top 20 in Ontario racing XC in the late 90s, and I’ve been on a bike pretty much my whole life. I’ve lost touch with where modern geometry has gone, I’ll own that. I’m no expert on the new numbers. But I’m also nobody’s fool when it comes to a bike, and “trust me” was all I got. No reasoning. No why. Just the weight of his experience landing on the table.
AI gave me the opposite. It walked me through long legs, a lower back that tightens up, hamstrings that do the same. It pointed out an old-school instinct where we size up because XC bikes used to be twitchy, a longer wheelbase calming things down on the trail. Then it told me modern bikes have already designed around that, so the old reflex might be sizing me out of the right fit. None of that is proof. But it’s reasoning I could actually hold up against my own situation, and a lot of it resonated.
So I’m sitting here with the salesperson’s conviction on one side and the machine’s reasoning on the other, and the honest truth is there’s no real advantage for either of them to be wrong. AI isn’t trying to sell me a size. The dealer might be running on a reflex that was right ten years ago. I’ve been misled by bike salespeople very recently, so I’m a little raw on the “trust me” front.
This is the pattern I keep running into in my actual work, and I’m only now finding the words for it. We are watching “trust me, I’ve been doing this forever” run straight into a world where anyone can get thorough, contextual, reasoned-out answers in thirty seconds. Experience is real. Intuition is real. But intuition that can’t explain itself is starting to look thin next to reasoning you can actually inspect. The expert who only has “trust me” left is in a tougher spot than they were a couple of years ago.
The funny thing is the one point everyone agrees on. Go ride it. The dealer says it, AI says it, and they’re both right. I haven’t ridden either yet (I wanted to get this down first so I wouldn’t bias myself). At some point a parking lot won’t cut it. I won’t know how it climbs, how it holds a line over jumps, how it actually feels under me on real dirt. At some point I have to trust something.
I genuinely don’t know where I’ll land. So I’m posting this before I decide, and I’ll follow up with what I actually chose and why. My money’s on the data resonating more than the certainty. But feel is a funny thing, and I’ve been wrong before. We’ll see.
The follow-up
Updated July 21, 2026
I said I’d close the loop, so here it is. I bought the extra large. The salesperson was right.
The setup was almost comical in hindsight. Every conversation had been quietly walking me toward the large. On paper it was a slam dunk. The reasoning held up, my own instincts lined up behind it once I heard it laid out, and it all made sense. I went in about as sure as I ever get.
Then I got on the bike. The large felt too small almost immediately. Not a close call, not a maybe. It just felt small under me, and it was genuinely strange, because everything on paper said it shouldn’t. I got on the extra large and it fit perfectly. Still compact, still quick, but right.
I keep thinking about this in Thinking, Fast and Slow terms. The salesperson ran on a thousand data points he’d long since stopped counting and handed me an answer with none of the math showing. Fast. The AI took every fact and reasoned it out in front of me, step by step. Slow. And the fast, can’t-explain-itself answer was the one that matched what my body told me the second I sat on the frame.
I want to be careful here, because it would be easy to turn this into human beats machine, and that is not the takeaway. Both of them told me to ride it. Both of them had real data behind their call. In the end it was my decision, and I sided with the salesperson because I think that was the right read. No hard feelings anywhere. AI, for its part, does not care. The salesperson got a little prickly when I mentioned I’d run it past AI, which I understand completely. Anyone would get a bit defensive.
The thing I keep coming back to is that his experience carried information that never made it into the numbers. He’d watched thousands of people my size sit on this exact frame and ride away happy. That is real knowledge, even when he can’t hand you the why for it. Human context and hard-won expertise are still enormously strong, and I think they always will be. The lesson isn’t that one side wins. It’s that the reasoning I could inspect and the experience I couldn’t both earned their place in the call, and I’m glad I had both in the room.
For the record, I bet on the data. The certainty won. I’m happy to say so.
Tired of Trust Me With No Why?
At Vervian, we help teams make calls they can actually inspect, where reasoning beats reflex. Let's talk about your next decision.