The Fly Easy today announced the launch of its new AI-powered price prediction engine, a machine learning system that analyses more than 15 years of historical fare data across 800-plus airlines to forecast optimal booking windows for travellers. The feature, available on both the website and mobile app, is designed to help users make more informed purchasing decisions and save an average of 23% per ticket.
The system, internally codenamed "FareSense," uses a combination of time-series forecasting models, real-time market signals, and airline-specific pricing patterns to generate a recommendation score for every flight searched on the platform. Users see a simple traffic-light indicator alongside search results: green suggests the current price is likely near its lowest point, yellow indicates the fare may drop in the coming days, and red warns that the price is expected to rise sharply.
"We built FareSense because we believe transparency is the most powerful tool we can give travellers," said CTO Laura Ducu, who led the development team from The Fly Easy's Bucharest engineering office. "Airline pricing is one of the most complex dynamic systems in commerce — fares can change multiple times per day based on demand, capacity, fuel costs, and dozens of other variables. Our models process all of that complexity and distil it into something immediately actionable for a traveller planning their trip."
The technology behind FareSense draws on several academic and proprietary advances in machine learning. The core model is a transformer-based architecture trained on a dataset of over 2.8 billion fare data points collected since 2009, covering every route served by the 800-plus airlines in The Fly Easy's search engine. This historical dataset is supplemented by real-time feeds from Global Distribution Systems (GDS), airline direct connect APIs, and third-party fare intelligence providers. The model is retrained weekly and validated against a holdout set of actual fare movements to ensure accuracy remains above 82% for seven-day forecasts.
In internal testing conducted during Q4 2025, FareSense correctly predicted fare direction (up or down) within a seven-day window 84% of the time, and within a fourteen-day window 79% of the time. The system proved particularly effective on high-volume routes such as Bucharest–London, Bucharest–Rome, and Cluj–Milan, where pricing volatility is highest and the potential savings are greatest. Users who followed FareSense recommendations during the beta period saved an average of $47 per round-trip ticket compared to those who did not.
The feature also includes a "Price Lock" integration that allows travellers to freeze a recommended fare for up to 24 hours while they finalise their plans. When FareSense indicates a fare is near its lowest, users are prompted to lock in the price with a single tap, protecting them against sudden increases while they confirm travel details with companions or employers. This combination of prediction and protection is, according to the company, an industry first at the OTA level.
Ducu noted that the team encountered significant technical challenges during development, particularly around the cold-start problem for new routes. "When an airline launches a new route, there is no historical data for us to train on," she explained. "We developed a transfer learning approach that leverages pricing patterns from analogous routes — similar distance, similar market dynamics, similar competitive landscape — to generate reliable predictions even before route-specific data becomes available. That was one of the harder engineering problems we solved."
The Fly Easy said it plans to expand FareSense's capabilities throughout 2026. The next phase will add seat-class-level predictions, allowing users to see whether upgrading to premium economy or business class is likely to become cheaper in the near future. The company is also developing an alert system that will proactively notify users when fares on their saved routes drop below predicted thresholds, eliminating the need to manually check prices. A hotel price prediction module is also in development, expected to launch in the second half of 2026.
Industry reaction to the launch was broadly positive. Seanvs Patel, chief innovation officer at Skyscanner, commented that "any tool that brings more transparency to airfare pricing is a win for consumers." Independent travel technology consultant Martin McBride described FareSense as "a genuine step forward in how OTAs can differentiate themselves through technology rather than just price wars."
FareSense is available immediately to all Fly Easy users at no additional cost. The company said that early adoption during the beta period exceeded expectations, with over 340,000 users enabling the feature within the first 48 hours of its general availability release. The Fly Easy expects the tool to become its most widely adopted product feature within the first quarter of 2026.