You've done this before. Why does it still feel new?
Say you've planned four or five trips using the same travel app or the same handful of browser tabs. You know you like a mid-morning flight over a red-eye. You know you'll pay more for a walkable neighbourhood and regret it if you don't. You know "boutique" in a hotel listing means something different depending on the city. None of that is a mystery to you anymore. And yet trip five starts exactly like trip one did: an empty search box, a blank itinerary, the same filters reset to default. Whatever the tool learned about you last time, if it learned anything, didn't come with you.
This has a name, and it's not a travel problem
Software that recommends things to people has a well-documented version of this called the cold start problem: a system "cannot draw any inferences for users or items about which it has not yet gathered sufficient information." A brand-new user gets generic, average results, because the system hasn't seen them do anything yet. That part is unavoidable: everyone starts as a stranger to a system once. What's not unavoidable is staying a stranger. The whole reason recommendation engines get better over time is that they keep what they learn. Collaborative filtering, the approach behind a lot of "you might also like" style suggestions, works on the basic premise that people who've agreed with your taste before will keep agreeing with it, which only works if the system remembers who agreed with whom. Take the memory away and you're permanently stuck at the cold start, no matter how many times you've used the thing. That's the quiet failure mode of most travel planning tools. Not that they're bad at trip one: plenty of them are genuinely fine at trip one. It's that they're just as bad at trip one the fifth time you do it, because nothing about trip four followed you into trip five.
Why travel tools stayed stuck at zero
Part of this is just how the category got built. Most travel products are optimized around the session, not the traveller: get someone in, get them to a booking or a click-through, done. There's not much product incentive to remember someone across sessions if the business model is one-and-done conversions rather than a relationship that spans years of trips. Part of it is also that "personalisation" in travel got redefined downward into something that doesn't require memory at all: more filters, more toggles, more "tell us your vibe" quizzes you fill out fresh every time. That's not memory, it's a slightly better-organised version of the blank page. It still asks you to re-derive your own preferences from scratch, it just asks more politely.
What actually compounding would look like
This is the part worth being precise about, because it's easy to overpromise. A system that genuinely remembers you doesn't need to guess less because it's smarter; it needs to guess less because it's already been told, and it didn't throw the answer away. The pace you like to travel at, what you count as a non-negotiable versus a nice-to-have, how "spontaneous" you actually are versus how spontaneous you say you are on a form: most of that is stable trip to trip. A system that keeps it doesn't have to ask again. That's the direction Journi is built toward, not fully realised yet. What's live today, still in closed beta, is Journi building an evolving understanding of how a specific person travels, their Travel DNA, and using that to shape the itinerary it hands back. The longer-term aim is for that understanding to compound across every future trip, so trip two draws on trip one automatically rather than needing to be told the same things twice. That compounding-memory version is where this is headed, not a claim about what's fully proven today.
The point isn't fewer questions. It's not being asked the same ones twice.
None of this is really an argument for a shorter intake form. Plenty of tools have tried that and just moved the cold start problem earlier without solving it. The actual difference is between a tool that gets marginally faster at asking you things, and one that eventually stops needing to ask at all, because trip four already told it what trip five needed to know. Right now, almost every travel tool on the market (Journi's current product included, honestly, while the memory layer is still being built out) asks you to reintroduce yourself more often than a relationship with five years of trips behind it should require. That gap between how well a tool could know you and how often it makes you explain yourself anyway is the actual problem worth solving, more than anything about interface or speed. If you've ever caught yourself re-explaining your own travel preferences to an app for the third or fourth time, that's the exact gap Journi is trying to close: not more options, but a system that keeps what it learns about how you travel and builds on it trip after trip.