The AI agent that booked my flights, and why I

I let an AI agent book a trip for me.

Here’s how it went. I told it: “I need to be in Austin March 15-18. I prefer window seats, morning flights, and hotels within walking distance of downtown. Budget around $800.”

Fourteen seconds later, it had an itinerary. Flights on Southwest. A boutique hotel on Congress Avenue. A restaurant recommendation for the first night based on reviews matching my stated preferences for quiet places with good whiskey selections.

The itinerary was good. Better than what I’d have found myself, honestly, because it cross-referenced flight prices across three days of flexibility I’d mentioned I had, and it found a Saturday departure that saved $140.

I was about to hit approve.

Then I stopped.

The pause

I hadn’t checked any of it myself. Not the airline. Not the hotel. Not the price. Not the location. Not the reviews. I was about to spend $800 on a machine’s judgment, and I hadn’t verified a single detail.

That’s not how trust is supposed to work. Trust is earned. You verify. Then you verify less. Then eventually you trust. It’s a process.

What happened here was the opposite. I trusted by default. The AI presented a polished itinerary with confidence, and my default response was to approve. The friction of checking felt like more work than the trust felt like risk.

That inversion scares me.

Why it scares me

Not because the AI was wrong. It wasn’t. The itinerary was great. The hotel was perfect. The flight times made sense.

It scares me because of what it says about how quickly we outsource judgment. The AI was good enough that I almost didn’t think. And “almost didn’t think” is a dangerous place to be when the decisions get bigger.

$800 for a trip is small. What about a medical decision? A financial investment? A legal question? If I default to trust on travel, where else will I default?

I think AI agents are going to be useful. Genuinely, practically useful. I think they’ll save time and find better options and handle logistics I don’t want to handle.

But I also think the ease of them will erode something I want to keep. The habit of checking. The muscle of skepticism. The practice of thinking for myself about things that matter.

What I did

I checked the flights. They were real. Good times, good price. I checked the hotel. Four stars, great location, within budget. I verified the restaurant. It exists, the reviews match what the AI said.

Everything was fine. But the checking mattered. Not because I caught an error. Because I reminded myself that I’m the one making the decision. The AI suggested. I decided.

I want to keep that distinction alive. I’m not sure the technology will make it easy.

Related thinking: