Last month in Chicago, a United Airlines flight destined for London was on standby, awaiting 13 connecting passengers from Costa Rica. Traditionally, such delays would necessitate rebooking passengers on different flights. However, thanks to an AI tool named ConnectionSaver, the plane managed to wait for both the passengers and their baggage while still arriving in London on schedule.
This advanced system sends notifications to delayed passengers as well as those who are already on board, providing status updates and explanations. Such AI-driven tools maximize every opportunity for passengers to catch their flights before the boarding gates close.
Minimizing Delays and Optimizing Routes
United Airlines isn’t alone in adopting AI technology—Alaska Airlines, American Airlines, and several other US carriers are also leveraging these innovations to improve travel experiences. According to Helane Becker, an aviation industry analyst with TD Cowen, these AI solutions help reduce operational costs and optimize performance, including fuel efficiency.
The New York Times reports that AI has been assisting Alaska Airlines’ dispatchers in planning more efficient flight paths since 2021. “Think of it like Google Maps, but for airspace,” explained Vikram Baskaran, Vice President of Information Technology at the airline.
Two hours prior to departure, the system analyzes weather conditions, no-fly zones, and scheduled commercial and private flights registered with the Federal Aviation Administration (FAA) to propose the most efficient route. Pasha Saleh, Director of Development at Alaska Airlines, emphasized that AI handles “an immense amount of information that the human brain can’t process.”
By 2023, roughly 25% of Alaska Airlines flights utilized this system, cutting a few minutes off flight times, which cumulatively added up to around 41,000 minutes and saved 1.8 billion liters of fuel.
Enhancing On-Ground Efficiency
On the ground, American Airlines and others have implemented the AI-powered Smart Gating system. This technology directs planes to the nearest available gates with the shortest taxi time. If the initially designated gate is occupied, the system quickly identifies the next best alternative.
This results in reduced wait times on the tarmac for passengers. The airline estimates saving 17 hours per day in taxi time and conserving more than 5 million liters of jet fuel annually.
Mohan noted that using AI as a virtual apron controller can cut taxi times by up to 20%, particularly at major airports.
According to the New York Times, generative AI technologies like ChatGPT and Gemini are also enhancing communication between airlines and passengers. Jason Birnbaum, Chief Information Officer at United Airlines, mentioned that the airline had struggled to effectively inform passengers about delays. The ability to generate and send accurate, context-specific messages to the right individuals for 5,000 flights per day is beyond human capacity but feasible for generative AI.
For instance, in January, United Airlines passengers traveling from San Francisco to Tucson received a message explaining their delay due to runway repairs at the San Francisco airport, which were limiting the number of arriving and departing flights.
Addressing Urgent Issues
A detailed explanation can significantly alleviate frustrations for impacted passengers. Jamie Larounis, a travel industry analyst, recalled receiving a message last summer that his Chicago-bound flight was delayed due to a storm and crew scheduling issues. “Receiving a specific reason for the delay made me feel like the airline had everything under control,” he said.
Generative AI excels at summarizing text, making it a powerful tool for reading customer emails. Last year, Alaska Airlines became one of the first to implement AI for efficiently managing customer messages. The system “reads” each email and summarizes the issues presented.
“We used to read and process requests sequentially,” said Baskaran. But now, the AI system helps prioritize urgent emails. For example, an immediate request related to an upcoming flight would be given precedence over complaints about a past flight.
This AI system also assists human staff in formulating their responses, such as deciding to send discount codes, and drafting replies to customers. “Humans still make the final decisions, but those decisions are now more informed,” Baskaran told the New York Times.
However, some limitations exist. AI doesn’t always provide accurate information. In 2022, an Air Canada chatbot erroneously promised a passenger funeral benefits for booking a full-price ticket to a family member’s funeral. The passenger subsequently filed a lawsuit, and while Air Canada argued that the bot was an independent entity responsible for its actions, the court ruled that the airline had to compensate approximately $800 for damages and fees.
