I’ve spent a lot of time searching for flights.
For myself. For friends. And through years of working on travel products and speaking with travelers.
The pattern is almost always the same.
You open a flight-search website with a relatively simple intention.
You need to fly from Yerevan to Paris for a meeting and arrive before the evening.
You want to spend a weekend in Rome, leaving after work on Friday and returning on Sunday without turning the trip into a tour of airports.
You are planning two weeks in Thailand and are willing to accept one connection—but not a 30-hour journey just to save $80.
Then the search begins.
Dates. Airports. Filters. Tabs. Baggage rules. Connections. Arrival times. Airline reviews. Another search to see whether leaving a day earlier would help. Another tab because one option is cheaper but lands at 2:40 in the morning.
Forty minutes later, you have more information.
But you are not necessarily closer to a decision.
That is the problem we decided to build Apricot around.
Flight search solved access. It did not solve the decision.
The travel industry has become exceptionally good at showing us flights.
Almost every route, airline, schedule and fare is available within seconds. Search engines process enormous inventories and return hundreds of technically valid combinations.
From an infrastructure perspective, this is an incredible achievement.
From a human perspective, it often feels like work.
The assumption is that more information leads to better decisions: more results, more filters, more sorting options, more flexibility.
Sometimes it does.
But after a certain point, more choice simply transfers the work of understanding the market to the traveler.
The search engine finds the flights.
You still have to decide which one fits your life.
And that is rarely as simple as choosing the cheapest or fastest option.
For a work trip from Yerevan to Berlin, arriving before an evening meeting may matter more than saving $70.
For a weekend in Rome, a Friday-night departure may be worth more than a cheaper Saturday-afternoon flight that removes half the trip.
For two weeks in Bangkok, one reasonable connection may be fine. For three days in Dubai, the same itinerary may make no sense.
There is no universally best flight.
There is only the flight that makes sense for you, for this trip, at this particular moment.
Most flight-search products were designed around inventory.
We want to build around intent.
Saving time means more than making search faster
When we say Apricot should save time, we do not mean returning results a few seconds faster.
Flight results are already fast.
The cost comes afterward: opening tabs, comparing imperfect options and repeating the same decision from slightly different angles.
The deeper cost is attention.
You can spend an hour searching and still feel that the task is unfinished. Maybe prices will change. Maybe another date is better. Maybe you missed an airline. Maybe you should check again tomorrow.
The uncertainty remains even after the browser is closed.
Something that should be the exciting beginning of a trip becomes administrative work.
I believe time is one of the most valuable things a product can return to a person.
Not because every minute needs to become productive.
The opposite.
People should spend less time operating software and more time on the parts of life that belong to them: family, friends, meaningful work, rest—or simply enjoying the anticipation of the trip.
A good product should not only help complete a task.
It should remove the need to keep carrying that task in your head.
That is the standard we want Apricot to reach.
Personalization should begin with understanding, not settings
Most products describe personalization as a list of preferences.
Preferred airlines. Seat class. Maximum stops. Home airport.
These details are useful, but real preferences are contextual.
You may normally choose the cheapest reasonable flight, but not when you need to reach Paris before a meeting.
You may dislike early departures, but accept one if it gives you another full day in Rome.
You may be comfortable with a connection on a long trip to Thailand, but not for a weekend in Istanbul.
The right choice changes because the trip changes.
Flight search should therefore feel less like configuring a database and more like explaining the situation to someone who understands you.
You should be able to say:
“I need to fly from Yerevan to Barcelona next month. I can leave Friday evening or Saturday morning. I’m traveling with a backpack, and I want to return late on Sunday so I don’t lose most of the weekend.”
That sentence contains more useful information than a dozen traditional filters.
It explains not only what is allowed, but what matters.
Apricot should notice those details, weigh the trade-offs and return a small number of flights that genuinely fit.
Not hundreds of options.
Three strong ones, each with a clear explanation.
One may protect more of the weekend.
Another may be meaningfully cheaper without making the journey painful.
A third may offer a better schedule and a safer connection.
The point is not to remove choice.
It is to make the decision understandable.
We do not want AI to produce more information
AI makes it easy to generate more.
More text. More recommendations. More combinations. More supposedly personalized content.
Travel does not need more output.
It needs better judgment.
The important capability is not the chat interface itself. It is the ability to take an unstructured human request and translate it into a useful decision.
“I don’t want to lose the whole Saturday” is really a constraint on departure and arrival time.
“I have a meeting that evening” means reliability matters—not only the scheduled landing time.
A preference for direct flights may be important when the price difference is small, but irrelevant when it is several hundred dollars.
A useful assistant has to understand those distinctions and explain the trade-offs clearly.
That is much harder than placing a conversational layer over existing search results.
It requires knowledge of the traveler, knowledge of the market and enough restraint to admit when the available options are poor.
Sometimes the right answer is to wait.
Sometimes it is to change one date.
Sometimes it is to pay slightly more for a flight that makes the whole trip better.
Intelligence is not only finding an answer.
It is recognizing which details should change it.
A useful alert should know which flight you would take
Price alerts already exist.
But most monitor routes and prices, not the traveler’s actual decision.
A route may technically become cheaper because a four-connection, 60-hour itinerary dropped by $10.
The alert is correct.
It is also useless.
It creates another notification to open, evaluate and dismiss.
We want Apricot to watch the flights that fit the real trip.
If the direct Friday-evening flight to Barcelona becomes meaningfully cheaper, that matters.
If the flight that gets you to Paris before your meeting becomes meaningfully cheaper, that matters.
If the price changes only for an itinerary you would never consider, you should not need to think about it.
An alert should bring you back only when something happens that could genuinely change your decision.
Why we are building it
We love to travel, and we believe people should have more time to do it—especially with the people they care about.
That is why we want Apricot to disappear into the background, quietly doing the work and giving travelers more time for the journeys, people and experiences that truly matter.
Finding the right flight is only the first step.
Our longer-term vision is to build a real travel friend: always available, familiar with how you travel and able to help with more of the journey as Apricot grows.
Not another tool that creates more work.
A companion that quietly removes it.