Air France-KLM said it has been able to block 70% of irrelevant
traffic to its system after implementing technology from Amadeus.
The “Advanced Airline Profile” filtering technology, which is
powered by machine learning, went live with two airlines
in May. It aims to help airlines manage search traffic as New Distribution Capability (NDC) volumes increase.
“This has enabled us to focus on the most relevant
NDC shopping requests, resulting in higher
conversion and lower system strain. That means sharper
performance for us, more meaningful queries for travel sellers and a smoother
experience for our travelers.” says Maxime Boussard, NDC program director at Air France-KLM.
Amadeus said the solution uses machine learning algorithms
to filter out irrelevant requests from travel sellers before querying the
airline’s system. It can also apply advanced parameters such as fare type and
day of the week.
The company said this means a significantly lower
look-to-book (LTB) ratio and less strain on airline systems.
Amadeus said the technology is now being deployed across the
airlines using its NDC solution.
“Every day, airlines are flooded with NDC shopping
requests—many irrelevant or simply unactionable. This leads to sub-optimal
processes for both airlines and travel sellers and represents a
significant cost issue for the whole industry. Amadeus Advanced Airline
Profile helps address this challenge by enabling NDC search at scale,” said
Delphine Domingues, head of travel distribution product marketing management at Amadeus.
In its third quarter earnings call, Amadeus president and
CEO Luis Maroto acknowledged that LTB ratios with NDC were already a challenge
in terms of transaction volume and something that could be exacerbated by artificial
intelligence (AI).
A recent article from PhocusWire explored the LTB and AI
issue as well as how the technology can be the remedy. It also discussed the
potential for charges to be levied for shopping.
Sabre also highlighted the issue recently, emphasizing the scale of the problem in its travel retail trends for 2026.
“In the 1990s, the average was around 10:1. By the end of 2025, it's
estimated it will have reached 1,000:1 or higher—and could soon reach
20,000:1 or even 200,000:1 as richer offers and personalization expand
exponentially,” the report said.
The company has said intelligent caching technology with AI to predict demand could be
the solution.