I was pretty disappointed to see that Disney Cruise Line is dropping their free Palo at brunch benefit for Platinum and Pearl Castaway Club members. The benefits of the loyalty program are getting pretty thin, which is a bummer. It also means I have to cancel my brunch for my next cruise (that I paid for months ago).
My family and I are 35 cruises deep, and as you might expect, they haven't all been perfect. Traveling with a kid that has ASD isn't always easy either (especially when I have it too). To that end, they've been amazing to us in some difficult situations. In person, they've always been kind and understanding. I've also had some deeper conversations via email with people that could genuinely create positive change. Their reputation is not by accident.
So when the Palo change came down today, naturally I had to register my frustration and disappointment. Sure, they now have a credit for dinner at Palo, but that comes at the expense of dinner with my kid, which I'm not skipping. I wasn't expecting my email to change anything, and I wasn't expecting anything in return. I did expect a human response.
Pretty sure that what I got was a robot.
Feedback like this used to take a few days. Now it came back almost instantly. If you've had colleagues send you AI mail, you know how the pattern: "I hear you, I'm empathetic, sorry." The first response even said that the benefit was never good for brunch. Either I got away with it onboard 25 times, or this was a classic AI hallucination making things up.
You can't fake hospitality. Feedback loops are what makes any product better, but in this space, it's essential. Disney knows this, or they wouldn't put so much emphasis on their surveys. You can't automate your way out of it, because AI still sucks at context, or even giving you observable facts. LLM's do not have wisdom or judgment, and can't separate minor things from serious problems. I fear that DCL is breaking a critical feedback loop.
I imagine things will be good on our upcoming cruise, but this interaction will likely become a failure point if they stick with it. Hopefully it's a sensible rollout where they're looking at the responses and flagging their inauthenticity and hallucinations.
No comments yet.