Skip to content
Journal
How It Works

Editing an AI itinerary by chat: the real test of a travel planner

August 2, 2026 5 min read

Generating an itinerary is the easy part. Any AI planner can fill three days with plausible stops when it’s starting from a blank page — there’s nothing to contradict yet, nothing booked, nobody tired.

The real test arrives later, usually mid-trip: it’s raining, or you slept badly, or the museum you built the afternoon around is unexpectedly closed. You need to change one thing. And the question that separates a useful planner from a pretty one is simple — can it change that one thing without breaking the rest of the day?

A day is a chain, not a list

An itinerary looks like a list, but a good one behaves like a chain. The 10 a.m. museum sits where it does because it’s a short walk from your hotel. Lunch is at 1 p.m. because that’s when the museum spits you out, and the restaurant is three streets away. The afternoon is gentle because the morning wasn’t.

Pull out one link and everything downstream shifts. The naive edit — delete the museum, insert something else that’s popular — ignores this. It leaves you with a lunch reservation forty minutes from wherever you now are, and an afternoon that no longer makes sense. A real edit doesn’t replace an item; it re-solves the day around your change.

Here’s what that looks like in practice, in four edits people actually make.

Edit one: “Swap the museum for something outdoors”

The setup: a Tuesday with a museum from 10 to 12:30, lunch booked at 1, a slow tram ride after.

The lazy answer is the city’s most famous park — across the river, nowhere near lunch. The good answer respects the chain: a botanical garden fifteen minutes from the restaurant, so the reservation survives. It also respects duration — a garden is ninety minutes, not two and a half hours, so the plan quietly adds a coffee stop to bridge the gap rather than leaving you standing outside a restaurant at 11:45.

One sentence in chat. Three small adjustments underneath. That’s the job.

Edit two: “It’s going to rain on Wednesday”

Weather is rarely a stop-level problem — it’s a day-level one. Grimly converting Wednesday’s outdoor plan into its indoor equivalents makes a worse Wednesday. The better move is usually a wholesale swap: Wednesday’s coastal walk trades places with Friday’s museums-and-covered-market day.

But a swap has fixed points that must not move. The dinner reservation on Wednesday night stays on Wednesday. And the swap only works if Friday’s museums are actually open on Wednesday — closure days quietly ruin more improvised plans than rain ever does. A planner that checks both before answering has done the part you’d otherwise do with six browser tabs.

Edit three: “We’re moving slower than planned”

Two days in, the pace that looked reasonable at home is clearly too much. So you say exactly that: make the rest of the trip lighter.

The tempting failure here is compression — the same eight stops squeezed into shorter visits, which is more exhausting, not less. The honest edit cuts. One anchor per day instead of two. Later starts. Lunches that are allowed to take ninety minutes. The single thing you’d genuinely regret missing stays; the things that were only there because the day had room go.

This is where a planner that asked about your energy in the first place has an advantage: it knows which stops were load-bearing and which were padding.

Edit four: “Make Thursday cheaper”

Budget edits sound simple and are quietly the most judgment-heavy. Anyone can strip a day down to free walks. The skill is knowing which expense is the day’s spine and which is decoration.

If Thursday exists because of the paid thermal bath, the bath stays — and the €60 restaurant becomes a market lunch, the ticketed viewpoint becomes the free hill with the same skyline. The day loses cost, not its point. A cheaper day that no longer has a reason to exist isn’t an edit; it’s a deletion with extra steps.

The measure of a good edit is what didn’t change

After any of these, look at what survived: the bookings, the day’s shape, the walking distances, the one thing you came for. A chat edit should feel like a small correction to a living plan, not a lottery where asking for one change risks getting back a stranger’s itinerary.

This is the standard worth holding any planner to, and it’s how editing works in Moodora: the itinerary is a draft you talk to. Say the thing in plain words — swap this, we’re tired, it’s raining — and the day gets re-solved around you, not rebuilt from scratch.

Plans should bend. The good ones bend without breaking.