Flight pricing is deliberately confusing. Airlines change fares constantly, the same seat costs different amounts depending on when and where you look, and the advice you find online is mostly recycled from 2016.
What has changed is that the tools now do the watching for you. Price prediction, flexible date searches, routing tricks no human would bother calculating. Below are the five worth your time, the prompts we actually use with chatbots, and the timing data, including the parts that are less reliable than they look.
What These Tools Are Doing
Underneath the marketing, they all do variations of four things:
- Chew through historical pricing, seat availability and demand patterns faster than any person could
- Forecast whether a fare is likely to rise or fall
- Spot pricing patterns that are not obvious, like a route that gets cheaper when a nearby airport has a sale
- Filter results around your preferences instead of showing everything
The honest summary is that they are good at monitoring and pattern-matching, and less magical than the marketing suggests. Used properly they save real money. They will not conjure a $200 fare to Tokyo.
The Five Tools Worth Using
1. Google Flights
The one to start with, every time. Fastest interface, broadest coverage, and the price graph alone justifies it.
It predicts whether a given fare is likely to go up or down, and on some routes offers a price guarantee that refunds the difference if the fare drops after you book.
How to use it properly:
- Enter your cities and dates, then ignore the results and click the date grid.
- Green dates are cheaper. Move your trip by a day or two and watch what happens.
- Scroll to the price graph to see the trend across a two month window.
- Click Track Prices and choose whether you want every change or only drops.
- If your destination is flexible, use Explore. Set your origin and dates, then browse the map by price.
Track more than one date range. The alerts cost nothing and each one is another chance to catch a drop.
2. Hopper
The prediction specialist. It analyzes an enormous volume of fares daily and tells you to book now or wait, claiming up to 95 percent accuracy and average savings around $50 per flight.
Treat vendor accuracy claims with the skepticism they deserve. The advice is still directionally useful.
The genuinely distinctive feature is Price Freeze. Pay a small deposit and the fare is held for up to seven days while you decide, which is the only tool on this list that solves the problem of finding a good deal at the wrong moment.
How to use it: download the app, enter your route, and read the color-coded calendar. Red is expensive, orange is average, green is cheap. Tap Watch This Trip for alerts, and check the When to Fly and When to Buy recommendations before committing.
3. Skyscanner
The one for people who do not know where they are going.
Type your departure city, set the destination to Everywhere, and pick Whole Month or Cheapest Month. You get a ranked list of everywhere you can afford to fly. It is the single best feature for anyone whose main constraint is budget rather than destination.
The multi-city search is also worth checking against booking each leg separately. Sometimes the combined fare wins, sometimes it does not, and there is no way to know without comparing.
4. Kiwi.com
The strange one, in a good way. Kiwi builds routes other sites will not, combining airlines that have no partnership and connecting through airports no sane itinerary would use.
The Nomad feature is the reason to use it. Give it a list of cities and it works out the cheapest order to visit them, which is a genuinely hard optimization problem to do by hand.
It also surfaces hidden city ticketing, where you book a longer itinerary and get off at the layover. Be careful. Airlines dislike it, it violates most carriers’ terms, and you cannot check a bag. Know what you are agreeing to.
5. Momondo
Best for the final comparison pass before you book.
The Flight Insight panel breaks down what is actually driving the price on your route, covering seasonality, demand and how much competition exists. It tells you whether a fare is high because of the date or because only one airline flies it.
Also check the Mix and Match results, which pair different airlines for each direction and frequently beat any single carrier’s round trip.
Using Chatbots for the Parts Search Engines Cannot Do
Search tools find fares. Chatbots are better at the reasoning around them, which airport to choose, which order to visit cities in, whether points beat cash.
Neither Claude nor ChatGPT can book anything, and neither has live fare data. Use them for strategy, then take the answer to a booking site.
Prompts we get real use out of:
Timing a trip. “What is the best month to visit Tokyo for cherry blossoms while avoiding peak flight prices? Weigh weather, crowds and typical fare patterns.”
Alternative airports. “Which airports near Paris are worth flying into instead of the obvious one? Compare distance from the center, transport cost and typical fare difference.”
Route order. “I want to visit Rome, Florence and Venice. What order is likely cheapest across flights and trains, and are there regional events that would affect it?”
Airline comparison. “Compare pricing, service and hidden fees for airlines flying New York to London, including budget carriers. Where do the cheap fares stop being cheap?”
Points versus cash. “For Chicago to Tokyo, how do I decide between miles and cash? Include any known sweet spots in award charts.”
The last one is worth pairing with Point.me, which searches award availability across loyalty programs and is far better at that specific job than a general chatbot.
Stacking the Tools
None of these is enough alone. The sequence that works:
Start broad in Google Flights to establish what the route normally costs. Check Hopper for whether to book now or wait. If your dates or destination are flexible, run Skyscanner’s Everywhere or Whole Month search. For anything multi-stop, try Kiwi’s Nomad. Do a final comparison in Momondo, checking Mix and Match before you commit.
Set alerts on more than one platform across several date ranges and nearby airports. Error fares get caught and corrected within hours, so the only way to benefit is to already be watching when one appears.
Three Ways People Actually Use This
The flexible traveler
Sarah wanted a trip the following week with no particular destination in mind. She asked a chatbot for five options from New York with good weather, interesting culture and cheap flights, then ran those through Skyscanner’s Everywhere search. Hopper said wait on Lisbon. She set a Google Flights alert, got a drop notification two days later, and booked $75 below what she first saw.
Total planning time, a couple of hours, for a city she had not considered.
The family trip
The Johnsons were going to Disney World from Chicago on fixed school dates, which is the hardest version of this problem.
They used Google Flights to find the cheapest dates in a three month window, set Hopper alerts on those dates and the ones either side, checked every Orlando-area airport including transport costs to the resort, and tested a multi-city itinerary flying into one airport and out of another.
That came to roughly $800 less than their first search.
The long trip
Mike wanted Tokyo, Bangkok, Istanbul, Rome and Rio in one journey. He asked a chatbot which order was likely cheapest given typical routes and seasonal events, ran the result through Kiwi’s Nomad feature, checked visa requirements separately, and used Flight Insight on each leg to time the individual bookings.
Around $1,500 less than booking each flight independently, which paid for an extra week of the trip.
Timing, and How Much to Trust It
These figures get quoted everywhere. They are averages across millions of fares, which means they are real and also nearly useless for your specific flight.
- Booking day. Tuesday is the one everyone cites, worth roughly 6 percent. This effect has weakened considerably as airlines moved to continuous automated pricing.
- Booking time. Between 3pm and midnight, up to 10 percent, with around 9pm often cited as the sweet spot.
- How far ahead. Domestic, one to three months out, 15 to 20 percent. International, two to eight months out, 20 to 30 percent. This is the one that genuinely matters.
- Flying day. Wednesday is typically cheapest domestically, about 8 percent. Internationally, weekdays beat weekends by 10 to 20 percent.
- Seasonal. Book summer travel in April or May for 12 to 15 percent. Holiday travel, three to six months ahead, 10 to 20 percent.
- Last minute. One to two weeks before a domestic flight can occasionally yield 20 to 30 percent, but only if you can walk away when it does not.
One number puts the rest in context. A given fare changes roughly 62 times over 11 months, moving about $36 each time. That churn is exactly why alerts beat trying to time it yourself.
Common Questions
Do these actually beat searching manually? For monitoring, yes, comfortably. Nobody checks a fare 62 times. For a single search on a route you know, the advantage is smaller than you would think.
Is it safe? The research tools are fine. Book through the airline or a reputable agency rather than anything unfamiliar. The risk is not the tool, it is some of the strategies, hidden city ticketing in particular.
How early should I start? Two to three months ahead domestically, five to eight internationally. Set the alerts early even if you are not ready to book.
Can a chatbot book for me? No. It can tell you which airport and when. You do the booking.
What is the catch? Predictions are probabilities, not promises. They miss. Always sanity-check layover times, baggage rules and which terminal you land in, because a cheap fare with a 55 minute connection in a large airport is not cheap.
The Short Version
Start in Google Flights. Be flexible on dates before you are flexible on destination. Set alerts and let them run. Book domestic one to three months out and international two to eight. Use a chatbot for the reasoning and a search engine for the fares.
That is most of the benefit, and it takes about ten minutes to set up.
Tried any of these? Tell us which ones earned their place and which wasted your time. For more on where AI is genuinely useful in daily life, check our AI Blog.
About the Author: Gavin Stacey
