I TRIED USING CHATGPT TO PLAN MY EDINBURGH TRIP - HERE'S WHAT HAPPENED

ChatGPT delivered specific Edinburgh recommendations—from Teardrop whisky bar in Leith to rainy-day routes through the National Museum of Scotland. But when it came to actually booking my flights, hotels, or even checking real prices, it hit a wall. After spending hours getting inspired by AI-generated itineraries, I still had to manually search dozens of booking sites, compare fluctuating prices, and handle reservations myself.

Here's what I learned about the gap between travel inspiration and actual booking—and why purpose-built travel AI beats general chatbots for real trip planning.

The inspiration phase: where ChatGPT shines

I started with a simple prompt: "Plan a 4-day Edinburgh itinerary for October, focusing on history and great food." ChatGPT delivered immediately.

The recommendations were genuinely impressive. It suggested starting with Arthur's Seat at sunrise (brilliant for photos), then warming up at Deacon's House Café on Brodie's Close afterward. For dinner, it recommended The Witchery by the Castle—atmospheric but touristy—alongside Fhior on North Castle Street for modern Scottish cuisine using Borders lamb and Orkney scallops.

ChatGPT excels at synthesizing travel knowledge, but it's essentially an incredibly well-read friend who can't actually call the restaurant for you.

The historical context was spot-on too. ChatGPT connected Edinburgh Castle's Stone of Destiny to the 1996 return from Westminster Abbey, making each site feel meaningful rather than just Instagram-worthy.

It even suggested practical routing: Royal Mile in the morning when crowds are lighter, National Museum of Scotland during potential afternoon rain, then evening drinks at Bramble on Queen Street. Smart sequencing that showed real understanding of how Edinburgh actually works.

Where the wheels came off: the booking reality

Energized by my AI-crafted itinerary, I moved to phase two: actually booking the trip. This is where ChatGPT's limitations became painfully clear.

"Find me flights from LAX to Edinburgh under $800" produced a helpful but useless response about checking "airline websites or travel booking platforms." No actual flights, no real prices, no availability.

I asked for hotel recommendations with rates. ChatGPT suggested The Scotsman Hotel on North Bridge and Hotel Missoni on George IV Bridge, but couldn't tell me if they had availability for my October 15-19 dates or what they'd actually cost. Every suggestion ended with "check their website for current rates."

The gap between travel inspiration and execution is where most trip planning falls apart—and where general AI reveals its fundamental limitations.

After two hours of manual searching across Expedia, Hotels.com, and individual airline sites, I'd found flights for $847 through British Airways with a London connection (not under $800) and discovered that The Scotsman Hotel was fully booked while Hotel Missoni wanted £280 per night—significantly more expensive than ChatGPT's vague pricing suggestions.

The AI had given me a beautiful framework with no foundation.

The memory problem nobody talks about

Here's what really frustrated me: ChatGPT couldn't remember my preferences between conversations. When I returned the next day to adjust my itinerary, it had zero context about our previous planning session.

I had to re-explain my $2,000 budget, October 15-19 travel dates, vegetarian dietary restrictions, and preference for boutique hotels over chains like Premier Inn. The AI treated me like a complete stranger, offering the same generic Edinburgh advice it gives everyone.

Real travel planning is iterative. You adjust dates based on flight prices, change hotels when your first choice is booked, modify activities based on weather forecasts. ChatGPT approached each query as a fresh start, making coherent trip evolution impossible.

The platform also couldn't learn from my booking choices. If I mentioned loving design-forward properties like The Bonham or preferring afternoon departures after 2 PM, that information vanished into the digital ether rather than informing future suggestions.

What purpose-built travel AI does differently

This experience made me appreciate why specialized travel AI exists. Services like Otherwhere understand that trip planning isn't about generating lists—it's about turning ideas into actual reservations.

When I described my Edinburgh vision to Otherwhere, I got back three curated flight options with real prices, real availability, and actual booking links. No "check airline websites" runaround. The service pulled live inventory from United, British Airways, and Virgin Atlantic and presented concrete choices I could immediately act on.

More importantly, it could hold my preferred British Airways flight departing LAX at 2:35 PM for 30 minutes while I decided—something impossible with general AI chatbots that can't access real booking systems.

The difference between travel inspiration and travel booking is the difference between reading about swimming and actually jumping in the pool.

The hotel recommendations came with verified availability for my exact October dates, real nightly rates including Edinburgh's tourist tax, and consideration of my stated preferences for character properties within walking distance of Princes Street.

Otherwhere also remembered everything. When I called back to adjust my departure date by one day, the agent knew my original preferences, my United MileagePlus number, and could instantly show how the change affected pricing across my entire booking.

The booking experience that actually worked

Instead of juggling multiple browser tabs and losing track of which option was which price, I texted my requirements and received three thoughtfully curated packages within an hour.

Option one prioritized value: British Airways economy departing LAX at 2:35 PM with a 2-hour London layover plus The Inn on the Mile on Royal Mile for $1,890 total. Option two balanced comfort and cost: United Polaris business class and The Caledonian on Princes Street for $2,340. Option three went full luxury: British Airways Club World plus The Balmoral at North Bridge for $4,200.

Each option included complete details: exact departure times (2:35 PM vs 11:15 AM), layover duration at Heathrow or Newark, hotel amenities like fitness centers and spa access, cancellation policies, and total pricing including Edinburgh's £2 per night tourist tax. No surprises or "starting from" asterisks.

When I chose option two, Otherwhere handled the entire booking process. I received United confirmation code HK7R9P, e-tickets, and Caledonian hotel vouchers directly—not affiliate links or "complete your booking" redirects.

The service also coordinated my United miles for the flights while booking The Caledonian separately to earn Marriott Bonvoy points. That kind of optimization requires understanding loyalty programs, not just generic travel advice.

The verdict: right tool for the right job

ChatGPT earned its place in my travel planning toolkit, but as inspiration, not execution. It's brilliant for discovering neighborhoods like Stockbridge, understanding cultural context around Edinburgh Festival Fringe, and generating creative itinerary ideas I'd never find on standard travel sites.

But for actual booking? General AI simply can't bridge the gap between recommendation and reservation. It lacks access to live inventory, can't remember your preferences, and fundamentally doesn't understand that travel planning is about making concrete choices, not generating infinite possibilities.

Purpose-built travel AI fills that execution gap by combining the inspiration capabilities of general AI with real booking power, persistent memory, and understanding of how travel actually works.

For my next trip to the Scottish Highlands, I'll likely use both: ChatGPT for creative brainstorming about Isle of Skye photography spots and cultural insights, then Otherwhere for turning those ideas into actual plane tickets and hotel keys.

Ready to skip the booking hassle on your next trip? Text (323) 922-4067 to get started with curated options that you can actually book, not just dream about.