Aviation Frontier

Qwen opened its “AI task-completion” capability beyond the Alibaba ecosystem for the first time, partnering with China Eastern to launch a full-journey AI ticketing service — described as the world’s most complete end-to-end AI air-ticketing flow. Passengers can search flights, book, select seats and check in via natural language, with official refund/change support, and future integration with membership, hotels and ride-hailing.

• Natural-language flight search & booking
• Online seat selection & check-in
• Official refund & change support
• Linked with hotels, maps and ride-hailing

Korean Air brought generative AI to its online chatbot, trained on massive data plus a knowledge base of aviation regulations and operating policies, with real-time intent recognition. Responses cite sources with links to reduce hallucination, support 13 languages, and seamlessly hand off to a human agent on request.

• Generative AI with real-time intent understanding
• Source-cited replies to reduce hallucination
• 13 languages supported
• One-tap handoff to human agents

KLM is a pioneer in social customer care. Its Messenger/WhatsApp chatbot BlueBot (BB) answers flight status, baggage rules and gate questions around the clock, and handles complex tasks like check-in and sending digital boarding passes, freeing human agents for tougher issues.

• Multi-channel via Messenger / WhatsApp
• Flight status & gate info
• Check-in & digital boarding pass
• 24/7 instant responses

AirAsia’s AVA (AirAsia Virtual Assistant), powered by Ada, handles 100,000+ queries a day across bookings, flight status, changes/refunds, baggage and loyalty — across web, app and messaging channels in 50+ languages, measurably reducing average handling time.

• 100,000+ queries daily
• Booking, changes, baggage & loyalty in one
• Web, app & messaging channels
• 50+ languages

Cathay Pacific partnered with Hong Kong language-AI firm Fano Labs to add advanced conversational AI to its chatbot, giving more accurate replies on flight status, booking changes and online shopping — meeting passengers’ shift toward WhatsApp and WeChat for instant, digital support.

• Accurate conversational-AI replies
• Flight status, changes & shopping
• WhatsApp, WeChat & digital channels
• Multilingual incl. Cantonese

United built “Agent on Demand” on Amazon Lex, letting passengers video-chat with airport staff from their phone to change flights and rebook. It draws on reservation data to suggest alternative routes on a map, helping roughly a million passengers a year affected by disruptions that self-service can’t resolve.

• In-airport video chat with live agents
• Smart rebooking during disruptions
• Alternative routes shown on a map
• Built on Amazon Lex + reservation data

In late 2024 Air France-KLM signed a strategic partnership with Google Cloud to apply generative AI across customer service, operations and data analytics — automating replies, generating real-time flight info and personalizing communications, while the group retains full ownership and control of its data.

• Automated, personalized customer service
• Real-time flight information
• Operations analytics & route optimization
• Airline retains data ownership

Built on Unisound’s “Shanhai” large model, the AI digital concierge serves Xiamen Airport’s T3/T4 departure halls as a lifelike virtual human, offering flight queries, facility info, indoor navigation, travel tips, dining recommendations and one-tap assistance — alongside a new Smart Airport Joint Innovation Lab.

• Face-to-face virtual-human interaction
• Flight queries, indoor navigation & facilities
• Travel tips & dining recommendations
• One-tap staff call

China’s first airport digital concierge powered by a large model launched at Fuzhou Airport, built by Unisound, Zhaoxiang and Huawei. Passengers simply ask “when does my flight depart?” or “how do I get to the gate?”; the lip-synced digital human serves 24/7 and learns preferences over time.

• China’s first LLM-powered airport concierge
• Voice queries for flights, gates & ID issues
• Real-time lip-synced digital human
• 24/7 service

Shanghai Airport Authority, Huawei and Fudian built a joint “Airport Brain & AI Innovation Lab”, driving a smart airport via an Airport Brain and digital twin. PVG deploys airside dynamic sensing, digital taxi guidance and high-precision vehicle positioning with AI-analyzed turnaround nodes; Hongqiao piloted China’s first civil 5G-AeroMACS, exploring aircraft path prediction, passenger-density detection and orphan-baggage recognition.

• Airport Brain & digital twin, full-scope situational awareness
• Airside dynamic sensing & digital taxi guidance
• Aircraft path prediction & passenger-density detection
• 5G-AeroMACS apronside trial

SITA launched cloud-based AI baggage analytics (SITA Bag Radar), turning baggage ops data into actionable intelligence with end-to-end visibility; its WorldTracer Auto Reflight auto-rebooks bags on the next best flight during disruptions. Integrating Apple Find My and Google Find Hub item-sharing cut permanently-lost bags by 90% and recovery time by 26% in year one.

• AI predicts baggage-disruption risks
• Auto-reflight of misconnecting baggage
• Apple Find My / Google Find Hub integration
• 90% fewer permanently-lost bags

Fuzhou Airport built a security-service-operations digital system: an AI-monitored departure-screen control (saving 35,000 kWh/yr, cutting troubleshooting by 43 min); an AI anti-intrusion system with 97% alarm accuracy; a boarding-bridge access-control system with facial recognition and video linkage; and service delivery robots that return lost items and handle courier delivery, lifting satisfaction by 15%.

• AI lossless real-time surveillance & anti-intrusion
• 97% alarm accuracy
• Facial-recognition bridge access control
• Service delivery robots

Delta combines real-time flight data, baggage location and connection times to have AI route baggage handlers optimally, lifting transfer efficiency by nearly 30%. Its Baggage AI system measurably improves overall handling. Meanwhile Delta Concierge began beta in the app, giving SkyMiles members real-time, personalized rebooking — pushing NPS up 25+ points during disruptions.

• AI-optimized baggage routes
• ~30% transfer-efficiency gain
• Delta Concierge personalized rebooking
• +25-pt NPS during disruptions

China’s Ministry of Transport’s “Ten scenarios of AI+Transport” highlights smart civil aviation: IoT monitoring and predictive-maintenance models for runways and bridges with early warnings and repair plans, airport-wide sensing and intelligent decision systems, AI-based aircraft/cargo loading, and proactive detection of pavement anomalies, runway incursions, drones and bird strikes with smart response.

• Runway/bridge health & predictive maintenance
• Airport-wide sensing & intelligent decision
• Uncrewed aircraft & cargo loading
• Proactive bird-strike / drone / pavement control

TravelSky had 21 achievements selected as typical airport-AI application scenarios, spanning AI+Safety / Operations / Service / Management: Xiamen Kaiya’s Ascend-AI full-process security screening, multimodal AI-visual perimeter-intrusion alerts, 3D-vision baggage handling robots, an airport smart-Q&A & report agent, and intelligent resource scheduling — aligning with national “AI + civil aviation” policy.

• 21 cases recognized as airport-AI applications
• Ascend-AI full-process screening
• 3D-vision baggage moving robots
• Airport smart-Q&A & report agent

IBM envisions an “AI-powered intelligent airport” where goal-driven agents automatically reallocate gates, turnaround and baggage resources during delays; Jakarta Airport deployed two AI biometric corridors for touchless passage, Miami uses biometric “face pods” for fast ID verification — aligning with IATA One ID for a “borderless” passenger experience.

• Goal-driven agents reallocate gates dynamically
• AI biometric corridors (touchless)
• Facial-recognition ID verification
• Aligns with IATA One ID

At Airport Show 2025, IDEMIA showcased multimodal biometric identity (instant face + iris recognition) and AI-powered baggage handling for future-ready airports — cutting processing time and congestion while keeping journeys safe and efficient, positioning digital identity as a core pillar of smart airports.

• Instant face + iris multimodal recognition
• AI-powered baggage handling
• Shorter processing, less congestion
• Future-ready airport digital identity

GE Aerospace teamed up with Microsoft and Accenture to launch its first generative-AI solution for aviation — a GenAI maintenance assistant that lets airlines and lessors access aircraft/engine maintenance records, lease details and document gaps in minutes instead of days/weeks; drawing on data from 44,000 in-service engines, it flags issues before failures occur. Carlyle Aviation was the first preview customer.

• Minute-level asset-record retrieval
• Generative intent query & reporting
• Predictive maintenance on fleet & engines
• Lease compliance & document-gap insights

Fliggy is pushing travel AI from “answerer” to “doer”: its AI trip planner generates personalized itineraries and completes bookings in one tap, while AI host “Nezha” runs 24/7 live-commerce. Fliggy stresses respecting the complexity of the real physical world to deliver AI that actually completes tasks across the whole trip.

• One-tap personalized itineraries
• 24/7 AI live-commerce host
• Completes tasks, not just Q&A
• Covers pre-trip to on-trip journey

Tongcheng’s self-developed agent DeepTrip generates personalized multi-theme itineraries that adapt mid-trip. For vague flight queries it parses intent and returns the best option within 5 seconds; it interacts in 17 languages for inbound travelers and ships as an open-ecology skill on major agent platforms with keyless deployment.

• Personalized itineraries that adapt mid-trip
• Parses vague flight queries in ~5s
• 17 languages for inbound tourism
• Open-ecology, keyless deployment

Booking.com’s AI travel assistant helps travelers search and book hotels, flights and other services in conversation, handles changes and cancellations, and recommends stays by destination, price and preference — a leading example of generative AI wired into a full booking funnel.

• Conversational search & booking
• Change & cancellation support
• Personalized stay recommendations
• Connected to huge global inventory

TraveLLaMA is a multimodal travel agent for complex planning that “sees” maps, street views and photos like a seasoned local guide. Using Travel-CoT reasoning to fuse visual and text data with constraints like opening hours and budget, it lifts itinerary accuracy — representing the multimodal frontier of T-AGENT research.

• Multimodal: reads maps, street view & photos
• Travel-CoT chain-of-thought reasoning
• Fuses opening hours & budget constraints
• Complex multi-day itinerary planning