We Can't Set the Fare. We Can Fix the Price Experience.
Xavier Mascaros · · 4 min read
Sunday night, Miami. A couple has a week off in April, $3,500 to spend and one rule: one checked bag each. They open a travel site and there it is, Cancún, flights from $140. Twelve tabs later they've learned three things. The $140 was one way, for one person, on one date. The bags cost extra on one airline and not on the other. And the hotel bundle does save money, but only once both pieces are already in the cart, and only on that site. Nobody lied to them. The price was just never designed to be understood.
That couple is who I kept picturing while working with a boutique U.S. travel agency this year. Their website looked like most small travel sites: four hand-picked deals, separate tabs for flights, hotels and cars, and a quote form at the bottom. Behind the form, an agent priced each piece in a different tool and emailed a quote a day or two later. By the time it landed, the fare had moved and the couple had booked somewhere else. The agency wasn't losing on price. It was losing on the experience of the price.
The more we mapped the traveler's journey, the clearer the pattern got. The fare is a snapshot of the date you typed, and every recommendation sticks to that single option, so the bundle math shows up after you've already committed. Nobody connects the dots between the flight, the stay, the car, the bags and the budget, let alone your actual situation: the days you have off, who's coming, the airline card already in your wallet. Add-on rules for bags, seats and changes vary by airline and fare and shift constantly, which makes bundles hard to price and the agency's own value nearly impossible to show. And under all of it sits the one thing no one in the room controls: airlines, hotels and cruise lines set the fares.
Look at the big names and each one solves a slice. Hopper made "buy or wait" mainstream, but its headline 95% accuracy figure comes from Hopper's own 2019 research, I couldn't find an independent audit, and the advice reads as certain without telling you how confident it is for your route. This July it also agreed to a $35M FTC settlement over pre-selected "Tip" and VIP Support charges, without admitting or denying the allegations. Airbnb now shows U.S. guests the full price with fees by default, but that arrived in April 2025 alongside the FTC's fee rule, and the total still stops at the stay. Expedia's packages carry real savings, yet the model depends on hiding the component prices, and Expedia's own guide admits the savings can vanish if you book the same flight and hotel separately. The deal only exists inside their walls. Google Flights is the best flight search there is, with a date grid, price insights and a filter that adds bag fees, but it's built for the average traveler. It doesn't know your status or your card, and its budget tool caps the flight, not the trip. It's a search engine, not a planner.
Meanwhile, the rule that would have forced airlines and agents to show bag and change fees at the first price display was vacated by the Fifth Circuit in February 2026. Showing travelers their real cost is a choice again. That makes it an advantage for whoever chooses it.
So we started from the constraint. We can't set the fare. What we can control is how the price is explained, when it's shown and what it adds up to, and that's exactly the kind of work AI is good at. We designed a travel manager the couple could talk to in one sentence. Instead of one fare for one date, it shows the whole trip's cost across seven nearby departure days, with a forecast that says how confident it is and why. It applies their bags, their seats and that airline card to every price before they see it, and every fee rule carries the date it was last checked. It shows the bundle math before anything goes into a cart, along with the one add-on that's actually worth adding. It fits the whole trip into their $3,500 and tells them what to book now and what to keep watching. And at the end it hands them a receipt of what the agency saved them, which finally makes the agency's value visible.
None of that works without guardrails, because a pricing tool that over-promises loses trust faster than a static page ever could. Forecasts show their confidence, never certainty. Nothing paid is ever pre-selected. The card is held until the supplier confirms, then charged. Staff prices override the machine, and any question the AI can't answer goes to a person.
In my last post I wrote about the pitch clock and how baseball kept the rules and changed the act: the grandfather, the grandson and the pitch clock. Travel pricing is the same game. The fares will keep moving, and no model is going to change that. What we can change is everything around the fare, so that the couple on a Sunday night in Miami sees the real trip on the first screen instead of the twelfth tab.
If you want to see how it works, the full case study walks through the problem, the journey, the teardown and the wireframes, and has a live demo where you set a budget, bags and dates and watch the price window and bundle math move: read the AI travel manager case study.