Item — HN Reader

I imagine this is a much harder problem than meets the eye. Even for my own travel, there are like a dozen things I'm trying to optimize for. Time, money, seat, travel to/from airport (traffic, train cost, taxi pickup/dropoff ease), lounge quality, layover length or complexity vs cost calculus, loyalty program, airport navigation, airport amenities, flight hospitality quality etc. etc.

When an AI can take into account all these subtleties and work reliably then yeh, I'd give it a go. But for now I'm kinda ~ok with the time cost of finding a non-rubbish flight.

I guess this is where 'travel agents' once used to shine.

Also: I want to be _absolutely_ certain the flight was booked -- that nothing was dropped, that no weird API failed to callback, that no letter was dropped from my name, that no silly email parser thought I meant London, Ontario, Canada.

Ps. Cool demo!

5 comments

> I guess this is where 'travel agents' once used to shine.

Score one for self-service software "improving lives" and "increasing productivity" by allowing everyone to do the job that would previously be done much more efficiently by dedicated specialists :).

Still, reading that:

> Even for my own travel, there are like a dozen things I'm trying to optimize for. Time, money, seat, travel to/from airport (traffic, train cost, taxi pickup/dropoff ease), lounge quality, layover length or complexity vs cost calculus, loyalty program, airport navigation, airport amenities, flight hospitality quality etc. etc.

I envy your dedication. Halfway through reading this list, I already feel like going by a train instead. I can't imagine doing it myself without some specialized software; I'd lose my mind trying to do it with the normal travel/accommodation search and booking websites.

> Also: I want to be _absolutely_ certain the flight was booked -- that nothing was dropped, that no weird API failed to callback, that no letter was dropped from my name, that no silly email parser thought I meant London, Ontario, Canada.

Very much this. It's my main concern here, too. I need live feedback on every input at every step throughout the process, because I don't trust software to not screw this up.

Thanks. Definitely more complicate than originally imagined.

A lot of this can be done automatically with time (eg, already built an auto-seat algo), but there is a reason the beta is focused on frequent fliers solving logistics problems and making repeatable decisions :) That said, think you'll be surprised by how well it optimizes already.

During beta onboarding, you make an account with all travel info (no need to parse this). Travelers also confirm their booking before purchase ("checkout" links to a page with all flight info in a lot more detail instead of 'access' link). And BonBook sends a confirmation email post-booking (required by federal regs).

I was very to booking a flight to SJO instead of SJC. The interface just said "San Jose" so it looked fine. Might have been nice to go to Costa Rica instead.

Checkout link in beta provides details. But you can request both :)

> I imagine this is a much harder problem than meets the eye. Even for my own travel, there are like a dozen things I'm trying to optimize for.

But if you created "profiles" of how you might search with various weight settings, it seems like it could get reasonably close without having to re-enter all of the details each time.

Not to mention cabin luggage allowances (as well as the likelihood they will be honored and you won't be checking your cabin luggage at the gate, or stowing your bag in row 45 when you're in row 9). Even finding the (cabin/checked) luggage allowances on an airline's website can be a task for AI.