How to Use AI to Plan Any Trip in 10 Minutes

I spent three weeks planning a four-day trip to Portland last year. Three weeks. Seventeen open browser tabs. A shared Google Doc that my friends never looked at. Two arguments about whether we should rent a car.

This year I planned a week in Mexico City in about ten minutes using ChatGPT. Better restaurants. Better routing. A rain backup plan I never would have thought of. And nobody argued because the itinerary was so good there was nothing to argue about.

Here's exactly how I do it now.

The Setup Prompt

Most people open ChatGPT and type "plan a trip to Mexico City." Then they get a generic list of tourist attractions that reads like a Lonely Planet table of contents from 2019.

The trick is giving it constraints. AI gets dramatically better when you tell it what you actually care about.

Here's the prompt I used:

"I'm going to Mexico City for 7 days in April with 2 friends. We're 23, into good food (not touristy), street art, live music, and walking around interesting neighborhoods. Budget is mid-range, not backpacker but not luxury. We don't speak much Spanish. We don't want to spend more than 30 minutes in transit between activities. Build me a day-by-day itinerary with specific restaurant names, neighborhood recommendations, and time estimates."

That's it. One prompt. What comes back is shockingly usable.

Why This Works Better Than Google

When you Google "best restaurants Mexico City," you get listicles written by SEO farms. The top result is always some place that was great in 2021 and is now overrun. You have no way to filter by what you actually like.

When you tell an AI your specific constraints, age, vibe, budget, neighborhood preferences, it synthesizes all of that at once. It's like having a friend who's been there recently and knows your taste, except that friend has read every travel blog and forum post ever written.

The itinerary I got had restaurants in Roma Norte and Condesa that my foodie friend (who'd been to CDMX twice) hadn't heard of. It clustered activities by neighborhood so we weren't zigzagging across the city. It even suggested which days to do which things based on typical crowd patterns.

The Follow-Up Prompts That Make It Great

The first itinerary is a draft. The magic is in the follow-ups.

"What if it rains on day 3?" You get an indoor alternative plan that fits the same neighborhood and timing.

"We're going to be tired after the flight. Make day 1 low-key." It restructures so your first day is chill exploration near the hotel, not a packed museum crawl.

"One of us is vegetarian." Every restaurant recommendation gets re-evaluated. It flags which ones have great veggie options and which ones you should swap out.

"What should we not do that most tourists waste time on?" This is my favorite prompt. It told us to skip the Frida Kahlo museum (2-hour line, small space, better to see her work at the main art museum) and to skip the Pyramids of Teotihuacan day trip unless we wanted to spend 4 hours in traffic. We went to Xochimilco instead. Way better call.

The Booking Layer

AI can't book things for you yet (not well, anyway). But it can do the annoying research part.

I asked: "For each restaurant you recommended, give me the Google Maps link, whether I need a reservation, and the average price per person in USD."

Took about 30 seconds to generate. Now I had a spreadsheet-ready list. I spent maybe 15 minutes actually booking restaurants and saving locations to a shared Google Map.

Compare that to the old way: open Yelp, read 40 reviews, check if they take reservations, look up the address, save it somewhere, repeat 20 times.

Packing and Logistics

This is the most underrated part. I asked:

"What should I pack for Mexico City in April? I want to pack light, carry-on only. We'll be walking a lot and going out at night."

It gave me a packing list that was genuinely smart. Light layers for the altitude (Mexico City is at 7,300 feet, which I forgot). Comfortable walking shoes. A light rain jacket for afternoon showers in April. A nicer pair of shoes for the mezcal bars.

Then: "What do I need to know about getting from the airport to Roma Norte? Include costs and scam warnings."

This is where AI really beats Google. Instead of reading a 2,000-word blog post to find the one paragraph about airport taxis, you get the direct answer: use the official taxi booth inside the terminal, pay upfront, expect about 300 pesos to Roma Norte. Don't take taxis from the arrivals curb. Uber works but sometimes gets stuck in the airport pickup area.

The Template You Can Steal

Here's a general-purpose trip planning prompt you can copy and modify:

"Plan a [X]-day trip to [destination] for [number] people in [month]. We're [ages], into [interests]. Budget is [budget level]. We're staying in [neighborhood/area or 'suggest where to stay']. Build a day-by-day itinerary with specific place names, time estimates between activities, and restaurant recommendations for lunch and dinner each day. Avoid tourist traps."

Then follow up with:

• "What if the weather is bad on [day]?"
• "Any safety tips we should know?"
• "What's the best way to get from the airport to our hotel?"
• "Adjust for [dietary restriction]"
• "What are locals doing that tourists miss?"

It's Not Perfect

AI sometimes recommends restaurants that have closed. It can be confidently wrong about operating hours. It doesn't know about that construction project blocking the main road to your hotel.

So verify the specific details. Google Maps the restaurants to make sure they exist and are open. Check hours for museums. But the structure, the routing, the neighborhood logic, the vibe-matching? That's all solid, and that's the part that used to take you three weeks.

The Bigger Point

Trip planning used to be a skill. You had to be good at research, good at logistics, good at synthesizing information from dozens of sources. Some people enjoyed that. Most people just procrastinated until they were winging it.

Now the skill is knowing what to ask for. Giving an AI your real constraints and preferences and letting it do the synthesis. It takes ten minutes and the output is better than what most people produce in weeks.

I've used this for weekend trips, international flights, even just "what should we do in LA this Saturday." It works at every scale.

If you want to see more stuff like this or want help figuring out how to use AI for something specific in your life, check out readlaboratories.com/learn or shoot me an email at jake@readlaboratories.com. Always happy to help.