B36 perspective · Airline strategy
Where airline AI earns its return
Five carrier models. Seven AI investment areas.
The strategic choices that turn smarter tools into advantage.
The executive read
AI advantage follows the decisions each carrier controls.
01 · Carrier economics
Five familiar models create different economic priorities.
Network, low-cost, regional, leisure and cargo operators sell different promises. Route structure, revenue mix and contracts explain why the same AI tool can have different returns.
Full-service / network
Connect cities through hubs; sell several cabins and a managed journey. AI has to coordinate pricing, service and recovery across connecting itineraries.
Low-cost / ultra-low-cost
Sell a basic fare with paid extras; build economics around efficient flying. AI can improve both ticket and extras pricing, while handling service at high volume.
Regional
Connect smaller markets, either for a larger airline or on the carrier’s own account. Aircraft reliability benefits the operator directly; fuel benefits follow the contract’s cost allocation.
Leisure / charter
Concentrate flying around holiday demand; some groups also sell the complete package. Pricing and aircraft reliability can protect peak-season flying and, for integrated groups, holiday contribution.
Cargo
Move freight by air or sell time-defined delivery through an integrated air-and-road network. Freight pricing must fit usable aircraft capacity; express AI must improve the stage limiting end-to-end delivery.
02 · AI investment
Pricing and service AI already cross carrier boundaries.
Wizz reports €1.10 incremental ancillary revenue per passenger from machine learning. Ryanair’s deployed service handles around 120,000 daily chat interactions. These measure different outcomes: revenue and service volume.
Seven areas: pricing, service, ground and cargo, flight efficiency, operations control, maintenance, and AI foundations.
Wizz disclosure · Ryanair case03 · Carrier patterns
The economic reason matters more than a disclosure ranking.
Our sample covers 39 parent organisations. Low-cost pricing is present; regional programmes cover maintenance and fuel; leisure disclosures span five areas. Freight and express differ because express owns more of the delivery chain.
Disclosed sample programmes; not representative of the carrier population.
04 · Future advantage
Sell the full basket, complete the repair and agree who keeps the gain.
Combine fares and extras with service costs. Link defect alerts to parts, engineers and repair slots. Protect sold departures before counting extra flights. When a partner pays fuel, agree how efficiency gains are funded and shared.
For freight, connect quotes to usable lift. For express, improve the stage limiting end-to-end delivery. Measure retained contribution alongside local task performance.
01 · Carrier economics
The carrier’s revenue model explains where AI can earn a return.
A network airline sells a connected journey. A regional operator may sell flying capacity to another airline. Both fly aircraft, but they control different revenue and cost decisions.
Choose a carrier type to compare the revenue model, operating constraint and AI rationale.
Full-service / network
Connect cities through hubs; sell several cabins and a managed journey.
02 · AI investment
Seven AI investment areas span the customer, the flight and the systems behind both.
Explore what each area does, then open the programmes to see how carriers implement it and what results they report.
Ground and cargo
Reduce gate delays, catering waste and parcel-handling effort.
Flight efficiency
Reduce fuel burn through better routes and flight profiles.
Operations control
Replan aircraft and crews when schedules are disrupted.
Aircraft maintenance
Find defects earlier and prioritise preventive repairs.
Investment area 01 of 07
Pricing and offers
Predict demand and set ticket, optional-product or freight offers.
Wizz Air
Ticket and extras pricing
Machine-learning models price tickets and optional products, including bags, seats and bundles.
€1.10
Incremental ancillary revenue per passenger
Company-reported effect of machine learning introduced in fiscal 2022; disclosed in October 2024. Revenue, not profit.
Read the programme source
Lufthansa Group
Ticket and extras pricing
Dynamic pricing adjusts fares and optional offers across Lufthansa, SWISS, Austrian and Brussels Airlines.
One group programme across four brands
The shared programme counts once in the group-based evidence map.
Read the programme sourceInvestment area 02 of 07
Customer service
Answer enquiries and complete permitted booking or service actions.
Ryanair
Conversational customer support
An AI service layer handles chat at scale and sends unresolved cases to human agents.
80%
Chats contained without a human handoff
AWS deployment case. Containment measures human handoff; the result is not a verified resolution or cash-saving rate.
Read the programme source
Air India
Booking changes and refunds
Agentforce executes selected refund and name-correction workflows within approved service processes.
14 days to 4 hours
Internal refund-processing time
September 2026 disclosure. Refund time covers internal processing; bank processing time is unchanged.
Read the programme source
easyJet
Customer email processing
eResolve uses AI to process recurring customer emails and support the service team.
+39 points
Reported email-response productivity
FY2025 annual report. The issuer does not define the point scale; these are retained as reported points, without conversion to percentages.
Read the programme source
Scoot
Conversational booking support
MARVIE combines conversational assistance with permitted booking checks and amendments.
+19
Percentage-point improvement in chatbot customer satisfaction
FY2025/26 annual report; MARVIE chatbot result. The starting score and measurement window are not disclosed.
Read the programme sourceInvestment area 03 of 07
Ground and cargo
Improve gate assignment, catering, physical sorting and shipment handling.
American Airlines
Gate assignment optimisation
Smart Gating assigns gates to reduce aircraft taxi time and gate conflicts.
20%
Lower aircraft taxi time at Dallas–Fort Worth
Airline disclosure, November 2023. About two minutes per flight at this airport; other hubs have different reported results.
Read the programme source
FedEx
Robotic parcel sorting
An AI-powered sorting robot handles a defined physical sorting task at its Cologne air-network facility.
Up to 1,000
Small parcels processed per hour by the sorting robot
Cologne deployment, June 2025. Published task capacity for documents and parcels up to 4 kg; not a throughput improvement versus a baseline.
Read the programme source
KLM
Meal demand forecasting
TRAYS predicts who will board each flight so catering teams load meals against expected attendance.
63%
Less food wasted versus catering every booked passenger
Three-month analysis disclosed in February 2024. Comparison is with booking-based catering; flights catered from Schiphol are the reported operating scope.
Read the programme sourceInvestment area 04 of 07
Flight efficiency
Recommend flight routes and profiles using weather, airspace and aircraft data.
Alaska Airlines
Flight route recommendations
Flyways analyses weather, traffic and airspace to suggest feasible routing alternatives.
3–5%
Fuel savings on flights longer than four hours
Airline report, May 2024. This result applies to Alaska; Horizon’s use is included in the regional evidence map without a separate savings result.
Read the programme sourceWizz Air
Flight efficiency recommendations
Machine-learning models recommend flight speeds and altitudes and improve taxi-fuel planning using flight and weather data.
0.5–1%
Fuel and associated CO₂ reduction per flight in operational testing
Wizz reports testing across more than 10,000 flights. This is the stated test scope, distinct from its wider fuel-efficiency programme.
Read the programme sourceInvestment area 05 of 07
Operations control
Forecast operational needs and support crew, aircraft and disruption decisions.
easyJet
Crew and maintenance forecasting
Its operations centre uses AI for aircraft maintenance and crew-demand forecasts alongside operational information tools.
Forecasts support control-centre decisions
Information retrieval, prediction and schedule optimisation are distinct capabilities. No isolated financial result is disclosed in this source.
Read the programme sourceInvestment area 06 of 07
Aircraft maintenance
Find recurring defects and support diagnosis before aircraft downtime grows.
easyJet
Component failure prediction
Skywise predicts component faults so engineers can replace parts before a technical failure disrupts a flight.
31+
Flights reported to have avoided disruption in the 2018 trial
Historical trial: 85 aircraft and three failure types. easyJet’s counterfactual estimate; not a current fleet-wide reliability rate.
Read the programme sourceRepublic Airways
Recurring defect detection
Ascentia Repeaters applies natural-language processing to maintenance logs to identify recurring faults.
Maintenance and reliability application
Collins’ April 2025 launch describes Republic’s selection. This is maintenance, not flight-efficiency AI.
Read the programme source
TUI Airline
Recurring defect detection
AI analyses technical-log entries and detects repetitive defects so maintenance teams can act on recurring issues.
Technical-log analysis in airline maintenance
First-hand implementation case. The operational mechanism is documented; an isolated airline profit effect is not.
Read the programme sourceInvestment area 07 of 07
AI foundations
Build common data, integration and permission layers for enterprise AI.
Riyadh Air
Enterprise AI architecture
Its IBM collaboration sets out an enterprise technology architecture intended to support AI across the airline.
Enterprise architecture commitment
A declared architecture programme, not a completed airline-wide transformation.
Read the programme source
UPS
Enterprise analytics and AI
Enterprise analytics and AI tools support activity across the integrated delivery business.
Enterprise deployment across delivery operations
UPS’s operating scope includes the wider parcel network, not just flights.
Read the programme sourceLive indicates an operating deployment; selected or committed indicates an announced programme; pilot indicates a limited test. Reported results retain their programme scope. The matrix includes programmes beyond these illustrations.
03 · Carrier patterns
The same investment areas serve different carrier economics.
Explore disclosed programmes across the five carrier families. Cargo is split into freight and integrated express here because air-freight capacity and end-to-end parcel delivery create different decisions.
Click a cell to jump to the programme detail below.
Analysis based on disclosed sample programmes; not representative of the carrier population. Counts include live, selected/committed and pilot programmes. Colour measures this sample, not AI spending. A zero means no qualifying case retained.
Full-service / network · Pricing and offers
3 reviewed organisations disclose pricing and offers programmes.
The economic opportunity extends to pricing the full itinerary and limiting the cost of connecting disruption.
Programme illustration
Lufthansa Group
Ticket and extras pricing
Dynamic pricing adjusts fares and optional offers across Lufthansa, SWISS, Austrian and Brussels Airlines.
What the matrix means for strategy
Pricing AI must cover the complete passenger basket.
Network airlines and low-cost carriers both use pricing AI. Ryanair earns 32.1% of revenue from extras: optimising the fare alone leaves a substantial part of the sale outside the decision.
Regional fuel savings can accrue to the mainline partner.
The operator controls the aircraft, but a capacity-purchase partner often pays fuel. Aircraft reliability protects the operator’s flying commitments; fuel AI changes the case for partner funding or gain-sharing.
Sorting faster and delivering faster are different gains.
Express robots can reduce handling effort even without changing delivery time. Faster delivery requires improvement at a binding stage of the air, sorting and road chain; freight quoting instead has to fit the flight’s usable capacity.
04 · Future advantage · Bridge36 perspective
AI creates more value when the sale, the flight and the contract are designed together.
The next investments should improve the complete transaction, protect sold departures and align funding with the company that retains the benefit.
A higher fare can still leave less profit.
Price the ticket and extras against the cost of serving the passenger, including likely rebooking. Use service AI to complete bookings and refunds, with the identity checks and system access needed to execute them.
Track the business resultContribution per sale and cost per completed service case.
AI maintenance earns its return through fewer delayed or cancelled flights.
AI can flag a part likely to fail before an aircraft breaks down. Arrange the spare part, engineer and repair time to fix it before departure. Less downtime first helps the airline operate flights already booked. Extra flights still require crews, airport capacity and passenger demand.
Track the business resultAvoided cancellations, completed contract flights and repair cost.
The company flying the aircraft may not keep the saving.
Where a partner pays for fuel, fund the tool together and agree how to share the saving. Price freight against each departure’s available weight and volume. Express automation can cut handling cost directly. Faster delivery requires fixing the stage holding up the parcel.
Track the business resultSavings retained under the contract, freight contribution, delivery cost and punctuality.
The leadership decision
Set the next AI priority from the economics of your carrier.
Translate the three perspectives into the next investment for your carrier type and the result it should improve.
Full-service / network · Next AI priority
Price a connecting itinerary against the cost of serving and recovering the journey.
Combine fares, extras and expected rebooking costs for the whole itinerary; preserve profitable connections.
Economic test
Contribution per itinerary after incremental service and disruption cost.
Revisit this carrier’s economicsBridge36
Turn smarter airline decisions into a stronger business model.
Price the complete sale. Protect the departure. Agree who keeps the gain. Build the AI capabilities that connect these decisions.
Bridge36 works with leadership teams to turn strategic questions into investment choices and delivery priorities.
Discuss your AI strategy