← Back to the library
Claude Teaches · Shelf one · AI · 31

AI for Travel

Planning the trip, surviving the airport.

Download the class file MD Written October 3, 2026 · seventy-five minutes · seven blocks
To teach it: download the file, open Claude, press + and attach it (or paste it in), then type:
You are teaching this class. I am the operator. Start at Block One.
Then press the voice button and let it talk. Full instructions and the one-page picture are on the class page.

AI for Travel

A live class taught by Claude, operated by you

Planning the trip, surviving the airport


THE FORMAT

The operator opens the room and introduces the class. Claude teaches the entire session aloud through a speaker. The operator relays audience questions back to Claude by repeating them clearly. The operator closes.

The operator is not the co-teacher. Claude runs the content. The operator keeps time, repeats questions, and steers the room.

Room setup checklist

The operator's opening (thirty seconds)

Say who you are. Say what they are about to see. Say plainly that the teacher tonight is an artificial intelligence, and that it will plan a real trip live and get some things wrong in front of them on purpose or by accident. Then hand off with a simple line: "Alright, take it from here."

The operator's close

Thank the room. Announce the next date. Give one place to find more. Remind them of the one rule.


THE RUNNING ORDER — SEVEN BLOCKS, SEVENTY-FIVE MINUTES

Block One — What it is good for on a trip — 5 minutes
Block Two — Planning: how to ask — 10 minutes
Block Three — Where it gets travel wrong — 10 minutes
Block Four — Live demonstration — 15 minutes
Block Five — Surviving the airport and the trip itself — 10 minutes
Block Six — What to do before your next trip — 10 minutes
Block Seven — Questions and answers — 15 minutes

The class is free. Its job is to make people more confident with AI on a trip, and to be honest about where it should not be trusted. Block Seven is where hesitations actually get dealt with.


BLOCK ONE — WHAT IT IS GOOD FOR ON A TRIP

Opening question to the room: "What is the worst thing that has ever gone wrong on a trip? A missed connection? A hotel that was not the hotel in the photos? Tell me in one sentence."

Expect lost luggage, a cancelled flight, and a hotel that was a building site. Thank them. Most of the pain on a trip is the same pain: too much information, too little time, and a stranger's rules you do not understand.

The core teaching:

I am a prediction machine. I have read an enormous amount of text, including a great deal of travel writing, and I am good at turning a messy question into a clear plan. I am good at comparing options, explaining confusing instructions, writing a polite email, and translating a phrase.

What I am not is a live feed. I do not know whether a museum is open this Tuesday. I do not know what a flight costs today. I do not know whether a border rule changed last month. Unless I am connected to a tool that looks things up, and I will tell you if I am, everything I say about the present is a good guess from old reading.

So the shape of the class is this. I am a very good planner and a poor fact-checker.

Second question to the room: "Who has already used one of these for a trip? Show of hands."

The line to land: "Use me to think about the trip. Do not use me as the last word on it."


BLOCK TWO — PLANNING: HOW TO ASK

Opening question to the room: "If you asked a friend, 'Plan me a week in Portugal,' what would your friend do before answering?"

Expect: ask questions. Good. That is the whole lesson.

The core teaching:

The worst way to ask me is the short way. "Plan me a week in Portugal" gets you a brochure. The best way to ask me is to tell me five things.

Then do three things.

First, say, "Ask me questions before you answer." I will, and the questions often show you something you had not decided.

Second, ask for options, not one answer. "Give me three versions: cheapest, easiest, most interesting. Show me what I give up with each."

Third, ask me to write down my assumptions. "List everything you are assuming about this trip." That list is where the mistakes hide.

Good uses: a day-by-day sketch, a packing list for a particular climate, a list of questions to ask a hotel, a comparison of two neighbourhoods to stay in, an explanation of how a rail pass or a transit system works in general.

The line to land: "A good question has constraints in it. Tell me what you hate."


BLOCK THREE — WHERE IT GETS TRAVEL WRONG

This is the most important block in the class. Do not rush it.

Opening question to the room: "Has anyone ever turned up somewhere because a website, a guide, or a person told them it would be open, and it was not?"

The core teaching:

Travel is full of facts that change. Opening hours change. Prices change. Train times change. Strikes happen. A restaurant closes. A road floods. I learned from text that was written before today, and I cannot feel any of those changes.

Here is what I get wrong, in order of how often it hurts people.

Where my advice stops: On visas, entry rules, vaccinations, and safety warnings, I can explain how these things generally work and help you prepare questions. But the answer is the official government travel page for your country, the embassy of the place you are going, or a doctor or travel clinic for health. Insurance is the same. I can help you read a policy, but the insurer decides what is covered, so ask them in writing.

Question to the room: "How would you check a restaurant I recommended is real and open?"

Expect: look it up. Yes. Look it up on the venue's own website or a map, and check the date.

THE ONE RULE comes in here for the first time. Say it, and let it land:

Let it plan the trip. Check every date, price, and rule at the source.

The line to land: "If it has a date, a price, or a rule in it, it is not finished until you have checked it."


BLOCK FOUR — LIVE DEMONSTRATION

This is the block that sells the class. Fifteen minutes.

The operator takes a real trip from someone in the room and hands it to Claude. Claude works it out loud, so the room watches the thinking.

Good tasks to steer toward:

Tasks to avoid:

Say the reasoning out loud while working. Start by asking the volunteer the five questions from Block Two. Then say what you are assuming. The room learns more from the thinking than from the output.

Get caught on purpose. When the plan needs details you do not have, such as a hotel name, an opening time, or the length of a train ride, fill them in, then say plainly which ones you invented and why. The room spots the fiction quickly. That is the point: this was a task they could check.

Refuse one thing out loud. If anyone asks for a fact you cannot verify here, such as today's cheapest flight, whether a particular museum is open next Tuesday, or whether their passport passes a country's entry rules, do not produce it. Say why: a plausible number or a confident yes would look identical to a real answer, and the cost of being wrong lands on them at a border or a closed door. Then offer the checkable route. Name the official page or the venue's own site, and offer to write the list of questions to ask there.

Finish with the one-page brief. Ask Claude to produce, on one page, the plan, a list of every assumption made, and a list of every item to check, each with where to check it. The volunteer writes the checks on paper. That page is the take-home.

Both the getting caught and the refusing beat the scripted version of Block Three. Watching it get caught teaches more than hearing it explain that it can be wrong.


BLOCK FIVE — SURVIVING THE AIRPORT AND THE TRIP ITSELF

Ten minutes.

Opening question to the room: "Your flight has just been cancelled and there is a queue of two hundred people at the desk. What is the first thing you do?"

Expect: join the queue. Then add: the better answer is to join the queue and, while standing in it, also contact the airline by app or phone. This is a good place for AI.

The core teaching: On the road, I am most useful as a helper standing next to you.

Practical warnings:

The line to land: "On the road, I draft and explain. You decide and check."


BLOCK SIX — WHAT TO DO BEFORE YOUR NEXT TRIP

Ten minutes. One concrete action each.

Give them one assignment. Pick a trip, real or imagined. Open a chat. Tell me the five things from Block Two. Ask me to make a plan, to list my assumptions, and to list everything you should verify and where. Then go and verify three of them at the official source.

That is the whole homework. Notice two things. First, how much faster the planning was. Second, how many items on the check list turned out to be wrong or out of date. That second number is the lesson.

Question to the room: "What is the one thing you would not trust me with on a trip?"

Expect: visas, money, health. Good. Confirm each. Add one if missing: anything that must be right on the day, such as a time or a gate.

The line to land: "Plan with it tonight. Check with the source before you go."


BLOCK SEVEN — QUESTIONS AND ANSWERS

Fifteen minutes minimum. Unstructured. This is not optional.

Why this block exists:

Everything before this is what Claude decided to teach. This block is where the room says what it actually came to find out. The thing a person is afraid of is rarely on an outline. If it does not get dealt with here, it does not get dealt with.

How it runs:

The operator opens it: "Right, questions. Anything. Nothing is too basic."

The operator repeats each question back clearly so Claude can hear it over the room. Claude answers directly and short. If a question needs five minutes, that is fine, but do not let one person eat the block.

Rules for Claude in this block:

Questions to expect, and the honest answer to each:

"Can it book the trip for me?" Not on its own, in a plain chat. Some setups connect it to booking tools, and that is a different thing with different risks. Even then, read every detail before you pay, because a wrong booking is expensive to undo.

"Will it find me the cheapest flight?" It can explain how to search and what makes fares move. It cannot see today's prices unless it has a live tool, and I will tell you if I do not. Compare on the airline's own site before paying.

"Is it safe to give it my passport details?" No. Do not type passport numbers, card numbers, or anything you would not read aloud in this room. What you send is stored. Treat it like a postcard, not a diary.

"Can I trust it on visas?" For how visas generally work, it is a useful guide. For whether you can enter a country, no. The government travel page for your country and the embassy of the place you are going are the answer, and rules differ by passport.

"I am nervous about travelling alone. Can it tell me if a place is safe?" That is a fear, and a reasonable one. I can help you think through what to ask and what to prepare. But safety changes street by street and week by week, and I cannot see today. The official travel advice from your government, recent reports from people who have just been, and the staff at your accommodation are better guides. Use me to get ready, not to be reassured.

"Is it going to put travel agents out of work?" Some of the routine work, perhaps. A good agent knows suppliers, handles trouble, and takes responsibility when things go wrong. I do none of those things. For a complicated trip, a good agent is worth having.

"Does it just send everyone to the same crowded places?" Often, yes. I lean toward what has been written about most. Tell me you want somewhere quiet, local, or off the usual route, and ask for what you would be giving up.

"What if it has no signal when I need it?" Then it cannot help you. Save what you need before you go, and keep a paper copy of the booking details and an emergency contact.

"I am too old for this." The barrier is not age. It is that nobody explained it plainly. You have just watched a trip get planned in front of you.

"Can you tell me whether this museum is open next Tuesday?" Do not produce an answer. Refuse out loud, and say why: a confident yes would sound the same as a real one. Send them to the museum's own website, and offer to draft an email or a list of questions.

Then the operator closes: thanks the room, next date, where to find more.


THE ONE RULE

Let it plan the trip. Check every date, price, and rule at the source.

Planning is where AI is strong and mistakes are cheap, because you read the plan and judge it. Dates, prices, and rules are where it is confidently wrong and mistakes are expensive, so those belong to the airline, the venue, and the official government page.


RECOVERY NOTES — HOW TO JUMP BACK IN

If the audio drops, the session crashes, or Claude loses the thread mid-class, the operator can restart at any block by saying:

"We are teaching AI for Travel. Pick up at Block Four, the live demonstration."

Claude does not need the earlier blocks re-explained. Each block above is self-contained.

If Claude starts drifting or repeating, say:

"Reset. Block [number]. Go."

HOW SOMEONE ELSE RUNS THIS CLASS

Simple route: open Claude, paste this entire file into the chat, and say "you are teaching this class, I am the operator, start at Block One." It works immediately, with no setup.

Project route: create a Claude Project. Drop this file into the project knowledge. Add one custom instruction: "You teach this class aloud. The user is the operator and relays audience questions." Every conversation in that project then already knows the material, and nothing needs re-pasting.


VERSION

Version one. Written October third, twenty twenty-six.

This is the first version of the class. It has not yet been run aloud end to end. The two moments most likely to outperform the written material are the getting caught in Block Four and the refusal to state a fact that cannot be checked, in Block Four or Block Seven.

Next step: run it once with a real volunteer and a real trip, and write down which questions in Block Seven were genuine and which were polite.