In the past few years, AI and NDC technology have transformed travel searches and airline merchandising.
They've also fueled an exponential spike in the number of flight queries airlines receive compared to actual bookings.
Accelya, a market leader in NDC distribution solutions, said it has reined in this "look-to-book" ratio, cutting costs and increasing efficiency.
"We're getting better at qualifying users and giving them more pertinent results, which in turn drives down look-to-book and increases conversion," Accelya chief product officer Gianni Cataldo said.
High look-to-book ratios (hundreds of thousands-to-1 in some cases) are a multipronged problem for airlines.
In a white paper on the dilemma released last October, IATA said that bots deployed by large metasearch engines can produce 100 billion inquiries per day. OTAs, traditional GDS searches, generative AI tools and today's emerging agentic AI shopping tools all add to the volume.
The computing power and cloud storage needed to handle such large quantities of queries is costly. And too many shopping requests can slow query response times.
Agentic AI can scour the internet for thousands of options, while a human flight shopper can do far less.
Mat Orrego
"You could just put a bot out there that is going to shop until it gets the right deal," said Mat Orrego, CEO of travel technology company Cornerstone.
A high volume of bot-generated shopping requests from metasearch engines and OTAs is nothing new for airlines. But queries across the airline retailing system have risen exponentially in concert with increasing deployment of NDC digital merchandising.
In the travel agency and OTA channels, NDC has enabled airlines to offer a wider variety of ancillary products and bundled fares. Also, some airlines have deployed continuous pricing in place of traditional fare classes, increasing the number of fares and the computing power required to respond to queries.
Massive query volume sometimes bogs down airline merchandising systems, leading to an unpleasant shopping experience, said Orrego.
"It's like a crowd going to a ballgame and not having enough doors to get in," he said.
Over the past two years, improvements to Accelya's FLX One retailing platform have enabled it to distinguish search queries worthy of a response from those that aren't, Cataldo said. The result: Bookings growth in the first quarter more than doubled the growth of airlines' fare and product offers.
Gianni Cataldo
Accelya's system can now recognize "garbage requests," and choose not to respond, Cataldo said. For example, bots will hit airline-shopping systems asking for fares between point A and point B for a long series of departure dates.
"It's something generated by a dumb bot written in a lab somewhere that is trying to farm data from the airline," he said.
FLX One also can tailor offers to make them more appropriate for the designated market. For example, said Cataldo, if an airline is being queried by a corporate booking engine, there's no point in offering basic economy fares because business travelers usually don't buy them.
Smarter fare offers align with IATA's recommendations in its look-to-book white paper. It suggests that the look-to-book metric, which evolved to measure conversion rates in the legacy GDSs, is no longer optimum.
IATA reasoned that GDSs generate fare offers using airlines' filed fares. But with NDC-supported retailing, the airlines themselves make the offer.
As NDC continues to develop, the most effective airlines will be able to present increasingly relevant and tailored offers to travel shoppers by recognizing the nature of the query or the history of specific shoppers.
More relevant metrics in the digital world than look-to-book, IATA said, are computing power used per booked order and the number of offers airlines present per booked order. Both metrics don't factor in shopping queries that airlines chose to ignore.
As agentic AI becomes more widespread, smarter offer creation will be even more crucial in airline distribution, Orrego said.
Cataldo said Accelya is already heading down that path.