There is a particular kind of holiday that looks, from the outside, like a fabulous success.
You went to Florence. You saw the David. You stood in the Piazza della Signoria at the hour when the light does the thing everyone says it does, ate at the restaurant four separate people had described as a hidden gem, and returned home with a camera roll that proved, beyond reasonable doubt, that you had been to Florence. The itinerary was completed. Every box was ticked. And somewhere on the flight home, you felt the faint, embarrassing suspicion that something was missing, though you could not quite say what.
This is the tyranny of the must-see. Not that the famous things disappoint, but that often people planning their holidays leave no room for anything else.
The must-see list has always existed in some form. Travellers of earlier centuries had their own version: a circuit of ruins, galleries, and cities considered essential for any person of education and taste. The difference is that those lists were compiled by people with genuine opinions, who had been to the places, who were willing to argue about them. They were idiosyncratic. They reflected something personal.
The modern must-see list is something else entirely. It is the aggregate output of millions of individual recommendations, averaged into consensus, stripped of context, and delivered at scale. It tells you where to go because everyone is going there, which is precisely the condition that makes going there least rewarding. And somewhere along the way, it stopped being a suggestion and became an obligation.
The must-see list found its natural home in the age of the internet, and it has found its logical conclusion in the age of artificial intelligence.
Ask any of the major AI tools to plan a week in Italy and it will do so with impressive speed and confident authority. It will suggest Rome for three days, Florence for two, perhaps a day trip to Siena and a final night in a Tuscan hill town whose name you recognise from a film. It will recommend the Colosseum, the Vatican, the Accademia. The itinerary will be, in every measurable sense, correct.
It will also be, in every sense that matters, a tyrannical must-see list. Because what the AI is doing — what it can only do — is synthesising the same vast pool of aggregated recommendations that produced the list in the first place. It has read everything written about Italy on the internet and returned the median answer. The median answer has nothing to do with you, your travel preferences, or your needs.
There is also the question of what happens after the itinerary appears on your screen. The AI has planned your trip. It has not booked anything. The hotels, the restaurants, the museum tickets, the transfers — all of that still requires you to open seventeen browser tabs and make decisions you may not even know how to make. People now spend more hours refining their prompts than they would have spent on a single conversation with someone who already knew the answers.
The deeper problem is not what the AI gets wrong, but what it cannot ask.
It cannot ask whether you have been to Italy before and, if so, what you loved and what left you cold. It cannot ask what you are recovering from or celebrating. It cannot notice that the way you described your trip suggests you would rather be somewhere quiet, or somewhere lively, or somewhere that feels nothing like your ordinary life. It cannot tell you that the famous restaurant is not worth it this season, but that there is somewhere else known only to the people who live nearby where the food is extraordinary, the room is calm, and you will feel, for the first time on the trip, like you have actually arrived somewhere.
The prompt is a question with a known answer. The best travel planning begins with questions that have no answer yet.
There has never been more travel information available, and it has never been less useful as a guide to what a trip should actually be.
Information tells you the opening hours of the museum, the rating of the restaurant, the photograph of the view. It is comprehensive and it is inert. It does not know you. It cannot tell you whether this particular place will mean something to you, whether this particular experience suits the kind of traveller you are, or whether the famous thing is worth the queue on this specific trip.
Knowledge is different. Knowledge comes from having been there, from having opinions formed by experience rather than aggregation, from understanding not just what a place is but what it is like on a Tuesday afternoon in October, who the right person to call is when you want a table that is not on the booking system, which of the well-reviewed options is genuinely worth it and which is coasting on a reputation it earned a decade ago.
This is the knowledge that good travel planning runs on. And it is exactly what the must-see list, by definition, cannot contain.
The solution is not to ignore the famous things. It is to build a trip around something larger: a sense of who you are as a traveller, what you actually want from the time, and what the place can offer when you are not trying to audit it. The must-see list is just the beginning. The rest is the point.
Tell us what you have in mind. A destination, a date, a feeling. We'll take it from there.
Begin a Conversation