I asked ChatGPT to plan a weekend in Orlando.
Not to find a hotel. To plan the trip. I told it I wanted to spend a day at Universal Studios with two park access, skip-the-line passes, and a hotel close enough to walk. I wanted to know my options and what they would actually cost.
What came back changed how I thought about the entire trip.
ChatGPT explained that three hotels connected directly to Universal offered something the math couldn’t ignore. Staying on property eliminated the need to buy a separate skip-the-line pass. The pass was included. The hotels were steps from the park entrance. For my dates, the total cost of staying slightly closer and paying slightly more for the room was lower than staying cheaper and buying the pass separately. For two people, the savings on passes alone exceeded $200.
I had not asked about hotels. I had asked about a trip. ChatGPT planned the trip and told me where to sleep as part of the answer.
I had one more question before I picked a property. I asked ChatGPT whether any of the three hotels had an Amex Platinum benefit. It said yes, identified one as the best option, and explained why. I went to the Amex Fine Hotels and Resorts site to see what the program was actually offering.
It was not a close call.
At roughly the same nightly rate the hotel was showing on its own website, Amex FHR presented a package that included a potential room upgrade, daily breakfast for two, a $100 resort credit, a welcome amenity, additional Amex travel benefits, and five times the points on the booking. Same price. A meaningfully different experience.
Before I booked I did what I suspect most people still do out of habit. I went to the hotel website.
The photography was generic. Shot selection that could have belonged to any large resort in any warm climate. No image communicated what made this property worth $500 a night or worth choosing over the other two options ChatGPT had given me. Finding the room rate for my specific dates required navigating through multiple pages with no clear path forward. The amenities were listed somewhere, eventually, after enough clicking. There was no single place where I could see what was included, what the room actually looked like, and what it would cost me to be there. The information I found in two seconds on a single Amex FHR page took me considerably longer to piece together on the hotel’s own site, and even then I wasn’t confident I had the full picture.
This was a large, well-known property charging over $500 a night. The website made it feel like an afterthought.
I went back to Amex FHR. At the point of booking, the choice was presented simply: $525 or 49,000 points. I had the points. The choice was sitting right there. I took it.
The hotel did not lose the booking. That is what makes this easy to miss.
The room was sold. The revenue was recorded. I arrived Sunday.
But the hotel did not explain the value. It did not frame the choice. It did not close the decision. ChatGPT made the trip make sense. Amex made the booking feel obvious. The hotel website made leaving feel rational.
What the hotel did not own was the decision path that produced the booking. Not the trip planning conversation. Not the pass economics that made their property the right answer. Not the moment I came to their site already sold and left because the experience gave me no reason to stay. Not the Amex relationship that assembled the offer. Not the points transaction that settled it. And not the guest relationship. Not really. Amex knew who I was, what benefit mattered, and how the transaction completed. The hotel received the reservation. It did not receive the path that created it.
What portion of those benefits the hotel absorbs, what Amex absorbs, and how the redemption settles are not visible to me. That is not the point.
The point is that a guest who was already decided, already at the website, already willing to pay full rate, left and booked somewhere else. Not because the hotel was the wrong choice. Because at the one moment they had a direct conversation with a guest, they were unprepared to have it.
A bad website is not a design problem. It is a failure at the only moment in the booking journey the hotel actually controls.
The hotel will never know it happened.
