How FairFares AI Finds Deals — Behind the Scenes
· By FairFares Team4 min readhow it worksAIdealstransparency

How FairFares AI Finds Deals — Behind the Scenes

TL;DR

There is a lot of marketing noise around "AI" in travel. Here is an honest, practical explanation of exactly how FairFares finds and scores flight deals — from data collection through to the alert in your inbox.

Table of Contents

🎯 Key Takeaways

ℹ️
✅ What you need to know
• FairFares fetches live flight prices across monitored routes every day — automatically
• Every deal is scored 1–10 using four factors: discount vs. route median, route popularity, seasonality, and days until departure
• The system does not predict prices — it compares current fares against real historical baselines
• Deals are ranked by score, not by discount percentage alone — context matters
• Alerts fire when a deal meets a meaningful score threshold, not every time a price changes

Here is an honest look at how FairFares actually works — because understanding the system makes you better at using it.

Step 1: Monitoring routes and collecting live prices

Every day, a scheduled job fetches current round-trip prices for each monitored route across upcoming departure windows. Routes tracked include combinations like AmsterdamLisbon, London Heathrow → Bangkok, and FrankfurtMálaga. For each route, the fetch captures the cheapest available fare, operating airline, journey time, stops, and a price range showing what is typical for that date.

That raw price, on its own, means almost nothing. Context is everything.

Step 2: Building historical baselines

FairFares builds route-level baselines over time: the median price for each origin-destination pair, a monthly seasonality index, and a sample count. When a new price comes in, the scoring engine compares it against the baseline for that specific route. A price 40% below the route median is treated very differently from a price 5% below — even if the absolute figure looks similar.

Step 3: The scoring algorithm

Every deal is scored 1–10 across four factors:

ScoreLabel
9–10Exceptional
7–8Great
5–6Good
Below 5Fair

Factor A — Discount vs. route median (40% weight): How far below typical is today's fare? A price at the median scores 0; a price 40%+ below scores 10.

Factor B — Route popularity (20% weight): Deals on well-documented routes are more trustworthy than routes with thin historical data. A route with 50+ samples scores 10; a new route scores 3 — an uncertainty penalty.

Factor C — Seasonality (20% weight): A cheap December fare to Málaga is more significant than a cheap September fare. If the departure is in a historically expensive month, the saving is harder to find and more meaningful.

Factor D — Days until departure (20% weight): The scoring sweet spot is 30–90 days out — far enough to plan, close enough that the deal is actionable. Very last-minute fares (under 14 days) score low because most travellers cannot act on them.

Step 4: Alerting

Once deals are scored, the system filters by threshold and routes alerts to subscribers via push notifications, daily email digest, and the Telegram channel. Alerts fire by score, not by price movement alone — a 5% price drop on a volatile route is not worth your attention. A 40% drop in peak season on a well-documented route, 60 days out, is.

This filtering is deliberate. Alert fatigue trains people to ignore alerts entirely. A meaningful score threshold means you hear from FairFares only when a deal actually clears a useful bar.

What the system does not do

It does not predict flight prices — it compares current prices against historical context. It does not cover every route — the monitored list is currently weighted toward popular European departures. It does not know your specific travel dates — deals correspond to departure windows the system happens to monitor.

A top-scoring deal is still worth checking for full fare conditions before booking, particularly baggage allowances on budget carriers.

Bottom line

FairFares monitors prices daily, compares them against historical baselines, and scores results across four factors — surfacing deals that are genuinely below the usual price, ranked by how meaningful the discount is in context. The "AI" in the description is accurate in that the scoring weighs multiple factors simultaneously. It is not accurate if it implies price prediction or personalised browsing learning.

Browse current deals on FairFares → — or join the Telegram channel for real-time alerts.

Share this article

By FairFares Team · Powered by ARAI