This article is published in Air Transport World part of Aviation Week Intelligence Network (AWIN), and is complimentary through Nov 09, 2024. For information on becoming an AWIN Member to access more content like this, click here.
Artificial intelligence (AI) in the air transport industry is not new. For decades, computers have crunched data in maintenance, repair and overhaul operations to identify and predict when parts need to be repaired or replaced before they fail, allowing airlines to get ahead of potential service disruptions.
AI-powered chatbots are now a common feature used within airline customer service, giving passengers the ability to retrieve their bookings, make changes and resolve simple queries. The ways airlines are using AI, however, are broadening and becoming more integrated with their wider operations. For example, Korean Air (KAL) is building an AI contact center in partnership with Amazon Web Services to consolidate customer care functions into a cloud-based platform, removing the need for physical space and servers. The center is expected to become operational in September, with machine learning and generative AI to be added by February 2025.
KAL told ATW that AI primarily will handle customer calls and chats, while human agents will remain available to offer personalized and advanced consultation services supplemented by AI technology. The airline said that compared with traditional customer service centers, the cloud-based AI system is expected to enhance agent responses, reduce operational costs, and make the use of resources more efficient.
The next step for airlines is to leverage generative AI in day-to-day operations. Alex Mans, CEO at FLYR, which uses generative AI in its real-time dynamic pricing and revenue forecasting systems, told ATW that the company filters canonical data models provided by airlines into a single data model that predicts future prices and revenue. FLYR also buys data from event organizers and online travel agencies, where the prices and revenue are re-forecast in real time when the context changes, such as the announcement of a major event like a concert.
The company’s goal for airlines is for 90% of prices to be set by algorithms and 10% by humans when, for example, there are major flight disruptions. FLYR is also using AI to manage and sell ancillary products based on an individual customer’s travel patterns and search filters.
Mans said airline data sets are separated and algorithms for each are not allowed to interact.
“Each airline is a bit unique. The customer base and product are slightly different, and they don’t observe the real time demand of each other, so there’s enough randomness to the system not to collude,” Mans said. “There will be scrutiny as to what’s different about AI—what’s improving or worsening both in choice and cost for the consumer. That’s going to be a big area of focus in the future, I think.”
CHINA FORGES AHEAD
In China, the potential of AI is both recognized and utilized. Airlines including China United Airlines, Juneyao Air, Loong Airlines, Ruili Airlines, Shenzhen Airlines and West Airlines are using Trip.com’s FlightAI revenue management system. Launched in 2020, FlightAI also features a market insight platform that is used by 50% of all Chinese airports and airlines to identify emerging trends and adjust strategies based on data derived from a wide range of sources in easy-to-understand dashboards.
Trip.com Group, China’s largest online travel agency, introduced TripGenie in February 2023, a generative-AI trip-planning tool that reacts to voice prompts via numerous languages including Chinese, Cantonese, English, French, German, Japanese, Korean, Spanish and Thai.
The tool can produce a computer generation of an itinerary, integration of hotel and flight bookings into a single platform, and personalized travel options. More importantly, Trip.com can finesse its AI engine simply through using data from social media, e-commerce and search engines. This vast amount of information enables the development of more sophisticated and accurate machine-learning models. The result is AI systems that are more efficient, capable of better pattern recognition, decision making and predictive analytics.
At a panel during the IATA AGM in Dubai, Mihir Shukla, founder and CEO of Automation Anywhere, an AI-powered software and automation company, said AI should be a “business strategy and not just an IT strategy” because of the impact it can have on an airline’s bottom line.
“Generative AI is good at managing unstructured data, but what’s new is how fast it is at doing it,” Shukla said, adding that reskilling is key to preventing push back from employees and to ensure they are not displaced.
IATA SVP information and data Kim Macauley, on the same panel, said AI should also be a data strategy because airlines need help in how they control and store data that interacts with multiple stakeholders, such as airports.
Macauley urged the industry to experiment with AI and attract technology talent into the system. “Think big and don’t hold back,” she said.
IATA director general Willie Walsh told ATW, “The pace at which we’re operating is incredibly fast and it’s clear that we believe AI can have positive and negative outcomes. I think everybody’s got to be careful about how they use AI because [we’re] going to have to address the ethical use of AI and that debate is one that’s ongoing, but work [in IATA] to understand all the facts has started.
“As we move forward with the use of AI, we will understand what the opportunities and the risks are associated with it to ensure that we’re not exposing ourselves or the industry to these risks.”
Pieter Elbers, chair of IATA board of governors and CEO at India-based carrier Indigo, added that he believed IATA would be able to work towards standardization on the AI front by working with industry and governments.
