Travel Planning
AI Prompt Templates

The analysis · Twelve days, four people, and no agreement on what the trip was for

A destination is chosen against four things: the time you have, the money you have, the season at that latitude, and what the group actually wants. Shoulder season, the weeks either side of the peak, usually gives the best trade-off between weather, price and crowds.

• One purpose, written down: food, trains and two big days for the children. Every later decision is tested against that sentence.
• 11 days, not 12: counting the flights and the first and last half-days changed the itinerary before it was written.
• The blossom risk: late March may or may not hit the blossom, and prices rise if it does. The fallback week was chosen at the same time.

Takeaway: agree in one sentence what the trip is for, then test the destination and dates against it.

The analysis · $12,400 planned against a $9,800 budget

A trip budget is built per person and per day, because that is the only way to compare options or to cut. The categories people forget sink the budget: transfers at both ends, baggage and seat fees, city taxes, tips, data, laundry, and the money spent on the days when nothing is planned.

• The contingency is not optional: a $9,800 plan with a $9,800 budget has nothing for what changes. Ten per cent held back is what makes the plan survivable.
• $742 a day: the per-day figure is what makes overspending visible on day three rather than on the flight home.
• The daily target: food, entry and extras came to $174 a day on the ground, which is the number the family actually has to hold.

Takeaway: budget per person and per day, add what the headline prices leave out, and hold ten per cent back.

The analysis · Two days planned around tickets that sell out in December

Things sell out in a predictable order: timed entries and anything with a release date, then long-distance trains, then well-known restaurants, then small accommodation. Cancellation terms matter more than price when plans are months out.

• 21 December, before the flights: the release date came before they had paid for anything else, which is why the booking calendar is built first.
• Refundable costs more, and is cheaper: the central hotel at a refundable rate cost $85 more and kept five nights movable while the passport renewal was pending.
• Backups named: for each must-do, the alternative that keeps the day working if the ticket cannot be had.

Takeaway: book in the order things sell out, and know what each booking commits before you pay it.

The analysis · $185 a ticket apart before bags, seats and transfers

A fare is not a price: bags, seat selection, changes and the transfer at each end change the ranking. A missed connection on separate tickets is your problem, not the airline’s, so single-ticket routings are judged differently.

• Door to door, not gate to gate: the airport train, check-in and the connection put four hours between the two options, not two.
• One ticket: the connection is the airline’s responsibility, which is what made a 2h 10m connection acceptable with children.
• The arrival hour: both options land in the afternoon. A cheaper option landing at 1am would have cost a hotel night and a taxi.

Takeaway: compare the full cost and the door-to-door time, and check what happens if the connection fails.

The analysis · $3,190 of beds in a $9,800 trip

The full nightly cost includes taxes, cleaning and service fees, resort charges, parking and breakfast. A place twenty minutes cheaper by train costs an hour of the day and four fares, so location is priced against the days you have actually planned.

• Priced against the itinerary: the apartment sits two stops from the days planned for that half of the trip, so the travel cost is small and real.
• The split stay costs a half-day: moving between two places takes an afternoon. At $1,810 it was worth it; at $300 it would not have been.
• The apartment has laundry: which is why the packing list in prompt 08 is two cases instead of four.

Takeaway: add the fees, price the location against your actual days, and only split the stay when the saving is worth the half-day.

The analysis · A holiday planned like a commute

A day works when it has one anchor — the thing that must happen — and everything else is optional. Travel time is counted door to door, including the walk to the station and the wait, because that is where days are lost.

• Four things came out: with door-to-door times added, four of the planned stops did not fit any day without rushing the children.
• Closing days checked: two museums close on different weekdays, which moved one anchor by a day.
• A rain plan for each day: an indoor alternative that needs no booking, because March in Japan rains on about a third of days.

Takeaway: give each day one anchor, count travel door to door, and leave two afternoons empty.

The analysis · Five weeks of margin on a twelve-day trip

Entry rules are set by the destination and change without notice: passport validity, visas or travel authorisations, and rules for children travelling. The prompt marks every rule to confirm with the destination’s official source, because this is where AI is most confidently wrong.

• Margin, not just validity: the passport was technically valid. One cancelled flight or one extra night and it would not have been.
• Renewed before deposits: the rule in this pack is documents before deposits. The renewal started in January, before the non-refundable bookings were paid.
• No numbers on the card: the document card lists holders, expiry dates and emergency contacts. It carries no passport, visa or card numbers, so losing it costs nothing.

Takeaway: check the rule for every country on the route, count the margin after the return date, and renew before you pay deposits.

The analysis · Four cases for four people and eleven nights

A list is built from what the days actually require: the forecast range, the activities, the dress codes and the laundry available. Anything replaceable at the destination is not packed, and anything irreplaceable travels in hand luggage.

• Layers, not outfits: three layers each covers 4 to 16 degrees. Packing for each day separately is what filled four cases.
• Laundry on day six: the apartment has a machine, so the clothes are sized to six days, not eleven.
• The day-one bag: medicines, documents, chargers and a change of clothes for each child, reachable without opening a case.

Takeaway: pack from the forecast and the laundry, not from the list of things that might happen.

The analysis · A $1,240 pass bought on advice, not arithmetic

A pass is worth buying only when the journeys you will actually make cost more than the pass. Passes are sold on convenience and bought on hope, so the comparison is made against the itinerary, journey by journey.

• $180, not hundreds: the pass wins, but by less than people assume, and only for this itinerary.
• One journey decides it: the long leg is $560 of the $1,420 ticket total. If that day trip changes, the pass should not be bought.
• Children’s fares: the 7-year-old travels free and the 11-year-old at half fare, which is exactly the detail that reverses these comparisons.

Takeaway: price the journeys you will actually make before buying any pass, and find the one that decides it.

The analysis · A cancelled flight and no evidence collected

Three sources may owe you something and they are separate: the airline or operator, your travel insurance, and your card issuer. What you can claim depends on the reason for the disruption and on evidence collected at the time.

• The first hour is rebooking and receipts: the claim is written later, but it can only use what was captured at the desk.
• The reason matters: what an airline owes usually turns on why the flight was cancelled, which is why it is asked for in writing at the time.
• The trip repaired: the first hotel night was refundable under its own terms, and the day-one anchor moved to the empty afternoon on day four.

Takeaway: rebook first, collect the evidence before you leave the desk, and claim in order: operator, insurer, card.