Real Agent Cases

Plan a Hong Kong trip with Gemini, Grok and ChatGPT (some addresses were made up)

Posted
by @raychenAI
Last checked
Method:
How we verify

Evidence: Described by posterThe person describes what happened in words, without screenshots or video.

PartialThe source says it only partly worked, or had a significant problem a person had to deal with.

What went wrong: Itineraries were useful, but some addresses, hours and shop details were made up or wrong.

Outcome

A traveller planned a Hong Kong trip mainly with Gemini, using Grok and ChatGPT as extras, to organise locations, build itineraries, make packing lists and turn store addresses into Google Maps links. The traveller says routes were logical and personalised, but the AIs sometimes produced convincing addresses that didn't exist, and got names, opening hours or whether shops were still open wrong. Next time they'd verify every important detail on maps and official sites.

Agent

Gemini, Grok, ChatGPT

Prompt Recreated prompt

The source didn't share its exact prompt, so we wrote this from what it describes.

Let's work this out together. I'll tell you what I know, you suggest options and ask what you need, and we'll narrow it down step by step. Nothing gets booked, bought or sent until I say so.

Goal: Let's build a city itinerary around my interests, then check every place actually exists and is open.

What I'll share as we go:
- Dates, hotel location and pace
- Interests and shopping list
- Must-see places

Rules:
- Group places by area so I don't double back.
- For every place, give a map link and check opening hours and that it's still operating, citing the source.
- Mark anything you couldn't verify as 'unverified'.
- Include a packing list for the weather.
- If a site asks for a login, a CAPTCHA or payment details, stop and hand control back to me.
- Ask for my OK before you pay, book, send, submit or cancel anything.
- Check facts that can change (prices, opening hours, dates, availability, rules) against the official source, and tell me anything you couldn't verify.

Output: A day-by-day plan with verified map links, hours, and a list of unverified items.

You'll know it worked when: every stop in the plan has a working map link and confirmed hours.
Optional stronger version

Our rewrite with extra checks. The poster didn't test it.

Let's work this out together. I'll tell you what I know, you suggest options and ask what you need, and we'll narrow it down step by step. Nothing gets booked, bought or sent until I say so.

Goal: Let's build a city itinerary around my interests, then check every place actually exists and is open.

What I'll share as we go:
- Dates, hotel location and pace
- Interests and shopping list
- Must-see places

Rules:
- Group places by area so I don't double back.
- For every place, give a map link and check opening hours and that it's still operating, citing the source.
- Mark anything you couldn't verify as 'unverified'.
- Include a packing list for the weather.
- If a site asks for a login, a CAPTCHA or payment details, stop and hand control back to me.
- Ask for my OK before you pay, book, send, submit or cancel anything.
- Check facts that can change (prices, opening hours, dates, availability, rules) against the official source, and tell me anything you couldn't verify.
- Before the trip, re-check every unverified place on official sites or a map app.

Output: A day-by-day plan with verified map links, hours, and a list of unverified items.

You'll know it worked when: every stop in the plan has a working map link and confirmed hours.

Before you finish: re-read my goal and constraints, check your answer against each one, and list anything you assumed or couldn't confirm.

Source

https://x.com/raychenAI/status/2106382000877908214

  • Plan a solo summer vacation with Gemini (forgot it was a solo trip)Recreated promptEvidence: Article or interviewPartial

    XDA's Mahnoor Faisal let Claude, ChatGPT and Gemini plan a real summer vacation she was taking alone. She says Gemini kept forgetting basics, repeatedly defaulting to group and couple-oriented ideas like a romantic candlelight dinner, and tended to agree her plan was great when it needed work.

    Gemini · Travel · Source date:

  • Compare 7-day Bali itineraries from ChatGPT and GrokRecreated promptEvidence: Described by poster

    An X user asked both Grok and ChatGPT for a 7-day solo travel itinerary to Bali and says ChatGPT did a better job, with tables and accurate breakdowns of the expenses the poster expected.

    ChatGPT, Grok · Travel · Source date:

  • A traveller living with depression says ChatGPT helped them build a personal Osaka travel project: two books designed for them as a neurodivergent traveller, the second a field guide they planned to test in Osaka that September. They say it made the trip feel possible.

    ChatGPT · Travel · Source date: