Planning a trip has a way of expanding. You open one tab about Lisbon, and four hours later you’ve got nineteen tabs, three conflicting opinions about the best day for the aquarium, and no booking. The planning becomes the holiday, and not in a good way.
AI is genuinely good here, as long as you use it for shaping and not for facts. It will happily invent an opening time. It won’t invent a sensible order to do things in.
Start with constraints, not destinations
The useful first ask isn’t “where should we go”. It’s “given all this, what actually works”.
Four days in Lisbon in late October, two adults and a nine-year-old. We’re staying near Alfama, no car, and we don’t want more than one big thing per day. The nine-year-old is fine with walking but hates queues. Give me a day-by-day shape, not a full itinerary: one anchor activity per day, a nearby backup if it rains, and where we’d realistically eat near each one.
“One anchor per day” is the whole trick. Lists of twelve attractions per day produce holidays where everyone is tired and someone cries, usually an adult.
Check the facts yourself, every time
Opening hours, ticket prices, whether the museum shuts on Mondays, whether that ferry still runs in winter. Treat all of it as a rumour until you’ve seen it on the official site. I’ve been caught out once and it cost us a morning.
The way round it is to ask for a checking list rather than answers.
For this plan, list everything I need to verify before booking, with the official source I should check for each. Flag anything likely to be seasonal or need booking more than a week ahead.
Twenty minutes with that list beats four hours of browsing, and you end up knowing which two things genuinely need booking now.
The packing and logistics bit nobody enjoys
This is where it saves real time, because the answer barely changes between trips.
Packing list for the trip above. Late October, mix of walking and one smart dinner, laundry available halfway. Group by person, note anything we’d need to buy, and keep it to what we’d actually use.
Save the output. Next trip, paste it back in and say what’s different. Two minutes instead of twenty, and you stop forgetting the adapter.
Settle arguments with a shortlist
Most travel disagreements are really about pace. One person wants three galleries, the other wants a long lunch. Put both in the prompt and ask for a compromise that names who gets what on which day. It sounds daft, but a plan that visibly gives everyone something keeps the peace better than any negotiation over dinner.
What to leave out
Passport numbers, booking references, card details, your home address. There’s never a reason to paste those in, and no useful suggestion depends on them.
A realistic time budget
Half an hour on the shape, twenty minutes verifying, ten minutes booking the two things that need booking. That’s it. The rest can happen once you’re there, where the decisions are more fun anyway, made over a coffee with actual weather in front of you.
Leaving gaps is the bit people find hardest. A day with one plan and an afternoon spare is usually the day everyone remembers.
Are you a planner or a wanderer when you travel, and does it match who you go with?