The short version, before the caveats
Yes, and it is genuinely good at part of it.
Ask ChatGPT for four days in Lisbon with a slow pace and a lot of food, and you will get a coherent, readable plan in twenty seconds. It is good at structure, at narrative, at explaining why one neighbourhood suits you better than another, and at questions like "is it worth staying in Alfama or Príncipe Real." That is real work, done fast.
The catch is that the parts it is bad at are the parts that cost you a day when they go wrong.
What it does well
**Shape.** Given constraints, it will produce a sensible arc for a trip. Four days, one of them a day trip, museums early, a slower final morning near your hotel. It reasons about pace properly.
**Narrowing.** "Which of these six neighbourhoods is quietest at night" is a question it answers well, because it is a question about general character rather than a specific fact that changes weekly.
**Drafting.** As a starting document for you to argue with, it is faster than any other method. Most people are better at reacting to a plan than making one.
**Explaining.** Why a place matters, what to expect, what the local convention is. This is where general knowledge is an advantage.
What it does badly
**It does not know what you saved.** This is the big one and it is structural. You have spent three months collecting places from creators you trust. ChatGPT has seen none of it. So it recommends the same places it would recommend to anyone, which means the plan it hands you is a generic plan, and the whole reason you were saving things was to avoid that.
You can paste your list in. Most people do not, because the list is forty screenshots and a folder of unlabelled reels, which is the original problem.
**It cannot show you a map.** A trip is a spatial problem. Whether two places sit together is the single most useful fact when ordering a day, and a chat window cannot answer it. It will tell you two places are close when they are forty minutes apart, because it is reasoning about names rather than looking at coordinates.
**Details go stale and it will not tell you.** Opening hours, closing days, whether a place still exists. A model's confidence does not drop when the fact is old. This is the failure that costs a day: you turn up on the one day of the week it is shut.
**It has no memory of the trip.** Ask about the same trip a week later and you start again, unless you keep the thread and scroll.
The pattern that works
Treat it as the ideas half and something else as the record half.
Use ChatGPT for shape, comparison and explanation. Which area to stay in, roughly how to sequence four days, what a neighbourhood is like at night, whether a day trip is worth it. All of that holds up.
Then verify anything time-sensitive against the source. Hours, closing days and whether a restaurant still exists are quick to check and expensive to get wrong.
And keep the actual places somewhere spatial, where each one carries the reason you wanted it and where you can see what sits near what.
Where Plotline fits
This is the part we built.
Plotline holds the places, not the conversation. You share a reel, a TikTok, a YouTube video or a Google Maps list, and every place in it lands on your own map with the creator's tip attached, so six months later the pin still says why. Then it plans the days around those places rather than around generic recommendations: destination, dates, pace, and it routes the ones you saved into an order that makes sense on a map.
The difference is not that one is smarter. It is that a chat window starts empty every time, and your saves are the thing worth starting from.
Using both together
The combination most people land on:
Ask ChatGPT the open questions. Which region, how many days, is this itinerary too packed, what is this neighbourhood like.
Save the specific places as you find them, from the creators who convinced you, with the context intact.
Let the planner do the spatial work, because ordering twelve places across four days by hand is tedious and a map does it better.
Verify the fragile facts before you go.
What we would not claim
We are not going to tell you AI cannot plan a trip. It can, and for a straightforward city break with no strong preferences it will produce something you could follow and enjoy.
What it cannot do is know that you saved a bakery in Graça because someone you trust said the custard tarts come out at ten, and that this is the reason you picked that morning. That is not a reasoning problem. It is a memory problem, and it is solved by keeping the places somewhere that remembers them.