AI in Passenger Airlines

Researchers at the SIA-NUS Digital Aviation Corporate Lab built LEAN-LLM-OPT, a three-agent LLM workflow that writes optimisation models from a plain-language query and data. On a booking simulator trained on Singapore Airlines sales data, it reached 93.33% and 92.06% execution accuracy on two revenue management tasks, ahead of Gemini 3 Pro and GPT-5.2.

Mirabel-based charter airline Nolinor Aviation partnered with Mila in late 2023 to apply an LLM plus an agentic retrieval layer to aviation safety reporting, reducing the human effort required to produce a full investigation report by up to 80% and cutting manual investigation time from 40 hours to 5 hours per case.

Delta Air Lines built its own AI system, working like a ride-sharing algorithm, to tell tug drivers at Hartsfield-Jackson Atlanta which transfer bags to move first and where to go. Delta says it has improved baggage transfer success rates by as much as 20% and plans to extend it to its Detroit and Minneapolis-Saint Paul hubs later in 2026.

Virgin Atlantic used OpenAI Codex to ship its revamped mobile app for the Christmas travel rush with near-complete unit test coverage and zero P1 defects at launch. Its engineers also report that Codex has cut legacy codebase size by 78 to 80% in some instances, and analyst teams now prototype internal applications directly on the company's data warehouse.

Air India has developed more than 30 AI tools in-house, from handling customer complaints to predictive decision making for maintenance of its 300-aircraft fleet. Its generative AI chat bot decides on refunds from customer emails and booking tickets, cutting call centre volumes by 50%, and the airline plans to save around Rs 100 crore annually.