Hacker News new | past | comments | ask | show | jobs | submit login

Booking a flight is actually task I cannot outsource to a human assistant, let alone AI. Maybe it's a third-world problem or just me being cheap, but there are heuristics involved when booking flights for a family trip or even just for myself.

Check the official website, compare pricing with aggregator, check other dates, check people's availability on cheap dates. Sometimes I only do the first step if the official price is reasonable (I travel 1-2x a month, so I have expectation "how much it should cost").

Don't get me started if I also consider which credit card to use for the points rewards.




Completely agree! Especially considering that flights for most people are still a large expense, people, especially those in the credit card points game, like to go to great lengths to score the cheapest possible flights.

For example, this person[0] could have simply booked a United flight from the United site for 15k points. Instead the person batch emailed Turkish Airlines booking offices, found the Thai office that was willing to make that booking but required bank transfers in Thai baht to pay taxes, made two more phone calls to Turkish Airlines to pay taxes with a credit card, and in the end only spent 7.5k points for the same trip on United.

This may be an extreme example, but it shows the amount of familiarity with the points system, the customer service phone tree and the actual rules to get cheap flights.

If AI can do all of that, it'd be useful. Otherwise I'll stick to manual booking.

[0]: https://frequentmiler.com/yes-you-can-still-book-united-flig...


Now THAT's the workflow I'd like to see AI agent automate, streamline and democratize for everybody.


If it were available to everybody, it would disappear. This is a market inefficiency that a "trader" with deep knowledge of the structure of this market was able to exploit. But if everyone started doing this, United/Turkish Airlines would see they were losing money and eliminate it. Similar to how airlines have tried to stop people exploiting "hidden cities."


This is not the same as, but reminded me of, patio11's writing on how banks have side channels for the professional-managerial class.

https://www.bitsaboutmoney.com/archive/seeing-like-a-bank/

> As a sophisticated user of the banking system, a useful skill to have is understanding whether the ultimate solution to an issue facing you is probably available to Tier Two or probably only available to a professional earning six figures a year. You can then route your queries to the bank to get in front of the appropriate person with the minimal amount of effort expended on making this happen.

> You might think bank would hate this, and aggressively direct people who discover side channels to Use The 1-800 Number That Is What It Is For. For better or worse, the side channels are not an accident. They are extremely intentionally designed. Accessing them often requires performance of being a professional-managerial class member or otherwise knowing some financial industry shibboleths. This is not accidental; that greatly cuts down on “misuse” of the side channels by that guy.


> Similar to how airlines have tried to stop people exploiting "hidden cities."

This sounds interesting?


https://skiplagged.com/

Just don't book a round trip, don't check a bag, don't do it too often. Also you're gambling that they don't cancel your flight and book you on a new one to the city you don't actually want to go to (that no longer connects via the hidden city). You can get half price tickets sometimes with this trick.


and watch it immediate evaporate or require even more esoteric knowledge of opaque systems?

Persistent mispricings can only exist if the cost of exploitation removes the benefit or constrains the population.


I have HAD a human assistant who booked flights for me. But it took them a long time to learn the nuances of my preferences enough to do it without a lot of back and forth. And even then, they still sometimes had to ask. Things like what time of day I prefer to fly based on what I had going on the day before or what I'll be doing after I land. What airlines I prefer based on which lounges I'd have access to, or what aircraft they fly. When I would opt for a connecting flight to get a better price vs. when I want nonstop regardless of cost. And on and on. Probably dozens of factors that might come into play in various combinations depending on where I'm going and why. And preferences that are hard to articulate, but make sense once understood.

With a really excellent human assistant who deeply understood my brain (at least the travel related parts of it), it was kind of nice. But even then there were times when I thought it would be easier and better to just do it myself. Maybe it's a failure of imagination, but I find it very hard to see the path from today's technology to an AI agent that I would trust enough to hand it off, and that would save enough time and hassle that I wouldn't prefer to just do it myself.


Off topic, but I’m curious: how did you go about finding an assistant that good?


I don't really need an AI agent to book flights for me (I just don't travel enough for it to be any burden) but aren't those arguments for an AI agent? If you just wanna book the next flight London to New York it isn't that hard. A few minutes of clicking.

But if you wanna find the cheapest way to get to A, compare different retailers, check multiple peoples availability, calculate effects of credit cards etc. It takes time. Aren't those things that could be automated with an agent that can find the cheapest flights, propose dates for it, check availability etc with multiple people via a messing app, calculate which credit card to use, etc?


In theory, yes. But in a real world evaluation would it pick better flights? I'd like to see evidence that it's able to find a better flight that maximizes this. Also the tricky part is how do you communicate how much I personally weight a shorter flight vs points on my preferred carrier vs having to leave for the airport at 5am vs 8am? I'm sure my answers would differ from wiradikusuma's answers.


Yep this is my vibe.

When I'm picking out a flight I'm looking at, among other things:

* Is the itinerary aggravatingly early or late

* Is the layover aggravatingly short or long

* Is the layover in an airport that sucks

* Is the flight on a carrier that sucks

* What does it cost

If you asked me to encode ahead of time the relative value of each of these dimensions I'd never be able to do it. Heck, the relative value to me isn't even constant over time. But show me five options and I can easily select between them. A clear case where search is more convenient than some agent doing it for me.


I agree. At first I would be open to an LLM suggested option to appear in the search UI. I would have to pick it the majority of the time for quite awhile for me to trust it enough to blindly book through it.

It's the same problem with Alexa. I don't trust it to blindly reorder me basic stuff when I have to shift through so many bad product listing on the Amazon marketplace.


Yep that's what I've been thinking. This shouldn't be that hard, at this point LLMs should already have all the 'rules' (e.g. credit card A buys flight X give you m point which can be converted into n miles) in their params or can easily query the web to get it out. Dev need to encode the whole thing into a decision mechanism and once executed ask LLM to chase down the specific path (e.g. bombard ticket office with emails).


And what happens to the 1% where this fails? At the moment the responsibility is on the person. If I incorrectly book my flight for date X, and I receive the itinerary and realise I chose the wrong month - then damn, I made a mistake and will have to rectify.

An LLM could organise flights with a lower error rate, however, when it goes wrong what is the recourse? I imagine it's anger and a self-promise never to use AI for this again.

*If you're saying that the AI just supplies suggestions then maybe it's useful. Though wouldn't people still be double checking everything anyway? Not sure how much effort this actually saves?


There is a really interesting book called Alchemy by Rory Sutherland.

In one chapter he describes his frustration with GPS based navigation apps. I thought it was similar to what you describe.

> If I am commuting home, I may prefer a slower route that avoids traffic jams. (Humans, unlike GPS devices, would rather keep moving slowly than get stuck in stop-start traffic.) GPS devices also have no notion of trade-offs, in particular relating to optimising ‘average, expected journey time’ and minimising ‘variance’ – the difference between the best and the worst journey time for a given route.

For instance, whenever I drive to the airport, I often ignore my GPS. This is because what I need when I’m catching a flight is not the fastest average journey, but the one with the lowest variance in journey time – the one with the ‘least-bad worst-case scenario’. The satnav always recommends that I travel there by motorway, whereas I mostly use the back roads.


The Flight Price to Tolerable Layover time ratio is something too personal for me to convey to an assistant


> Booking a flight is actually task I cannot outsource to a human assistant, let alone AI.

Because there is no "correct" flight. Your preference changes as you discover information about what's available at a given time and price.

The helpful AI assistant would present you with options, you'd choose what you prefer, it would refine the options, and so on, until you make your final selection. There would be no communication lag as there would be with a human assistant. That sounds very doable to me.


It is not that you can't outsource it, but there are so many variables that once you finished explaining them to the assistant (human or AI) you'd be better off doing it yourself. A human assistant only makes sense if he/she is your everything assistant and has knowledge about your work schedule, life, kids, financials, etc...


I feel the same way, or at least I wouldn't delegate this unless they fine tune accuracy and reliability in their apps. Right now, it sits around 40-60%




Consider applying for YC's Summer 2025 batch! Applications are open till May 13

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: