Airline pricing follows patterns. Most travelers don’t see them.#
If you’ve used Google Flights or Hopper, you’ve seen a version of price history — a chart showing how a fare has changed over the past few weeks. That’s useful but limited.
The more powerful signal is deeper historical data: how this route has priced over the past two or three years, across different seasons, booking windows, and market conditions.
That’s what LatAI indexes. And it changes what you can predict.
The booking window curve#
Every route has a characteristic price curve as a function of days before departure. It’s not the same for every route.
Long-haul international flights typically see a U-shaped curve: prices start high many months out (early birds pay a premium or get lucky), drop to a mid-range level in the 4–8 week window, then climb sharply as departure approaches.
Domestic short-haul routes in markets with high competition (e.g., the US Northeast Corridor, UK domestic) often have flatter curves with more unpredictable drops triggered by seat sales rather than time.
Last-minute international routes can go either way. Premium cabin fares often become cheaper 3–7 days before departure as airlines try to fill seats. Economy fares for popular leisure routes stay high or climb because leisure travelers booked months ago.
Knowing which curve applies to your route tells you whether to book now or wait.
Day-of-week patterns#
Across most routes, the cheapest days to fly are:
- Tuesday and Wednesday (lowest demand, mostly business-free)
- Saturday (business travelers go home Friday, leisure travelers leave Sunday)
The most expensive days:
- Sunday evening (business travelers returning)
- Friday afternoon (leisure travelers starting trips)
- Monday morning (business travelers starting weeks)
These patterns hold on most routes — but not all. Leisure-heavy routes (think: ski resorts, beach destinations) flip the pattern on weekends. LatAI identifies the day-of-week pricing pattern specific to each route.
Seasonal patterns by route type#
| Route type | Cheapest window | Most expensive window |
|---|---|---|
| Caribbean beaches | May–June, September | December–January |
| European city breaks | November–February | June–August |
| Southeast Asia | April–May, September | Chinese New Year, European summer |
| South America | March–April, October | December–January |
| Domestic US | Mid-January, mid-September | Thanksgiving, Christmas |
These are generalizations. LatAI looks at your specific route, not the category average.
What AI adds that charts can’t#
A price history chart shows you what happened. An AI model trained on millions of routes tells you what’s likely to happen next.
LatAI’s prediction model incorporates:
- Route-specific historical patterns (not global averages)
- Current demand signals (how fast are seats selling?)
- Competitive factors (are other airlines on this route running sales?)
- Macroeconomic factors (fuel price trends, seat capacity changes)
- News signals (events affecting demand)
The output is a Fare Prediction Score: not a specific price prediction (which would be overconfident), but a calibrated signal telling you whether the current price is historically high, average, or low — and which direction it’s likely to move.
The honest limits of prediction#
No model can predict airline pricing with certainty. Airlines have full discretion to change fares at any time, and they do. A sudden seat sale, a competitor going bankrupt, or an unexpected news event can blow up any prediction.
What historical data and AI models provide is probabilistic advantage — the ability to make better decisions more often than not. Over many trips, that compounds into significant savings.
Think of it like weather forecasting: not certain, but far better than guessing.
See how LatAI’s historical data engine works alongside our other signals.