Lufthansa lets AI draft the nightly maintenance plan for its fleet
Lufthansa says an AI planner now rebuilds the maintenance plan for its fleet of more than 300 aircraft every night, with human planners reviewing each work package before release. The group's other documented production use is a GPT-based enquiry bot at its pilot school.
The operating picture
Fitting the work into the ground time
Lufthansa Boeing 747-8 D-ABYI IAD VA1. Photograph: Acroterion / CC BY-SA 4.0 (cropped to fit). Illustrates the airline, not an AI deployment.
Every scheduled maintenance task has to fit into the time an aircraft spends on the ground, with the right parts, mechanics, tooling and hangar space available. Lufthansa told Porsche Consulting in April 2026 that an AI system now recalculates that plan for the whole fleet each night, grouping tasks that share an access panel and adjusting when aircraft are swapped at late notice.
The second record is smaller in scale. In October 2024 the group's European Flight Academy put George, which Lufthansa calls its first fully GPT-based bot, on the academy website to answer questions about pilot training. A third record, Amadeus passenger recovery for disrupted travellers, stays in the wider register: its attached sources describe automatic rebooking but make no claim that AI is involved.
Selected implementations
The work behind the evidence
01 / Deployed
A maintenance plan rebuilt every night
The operational problem
Planners must fit scheduled tasks into ground times across a fleet of more than 300 aircraft. The plan breaks whenever an aircraft is swapped, a part is late or hangar capacity changes.
What the evidence reports
Philipp Krusemeyer, senior director of digital TechOps at Lufthansa, said in an interview published by Porsche Consulting in April 2026 that the system recalculates the plan for the whole fleet nightly. It groups related tasks into work packages, accounts for parts, mechanics, tooling and hangar capacity, and passes the packages to the core maintenance system, where planners review them before production. Rollout began on a small part of the fleet. He said a planner can now schedule up to 30% more aircraft, and that the system finds about 15% more planning synergies than planners working alone.
Operational significance
Planners still sign off every package, which Krusemeyer described as a regulatory requirement. The gain is in planner capacity and plan quality, with the accountable human decision left in place.
Evidence still needed
Both figures come from a consultancy interview linked to an award nomination. "Up to 30%" is a ceiling, not an average, and neither figure has a published baseline, period or method. The interview does not say whether delays caused by maintenance have fallen.
A source can verify an announcement without verifying its projected outcome.
02 / Deployed
George answers prospective pilots
The operational problem
The European Flight Academy's commercial team handles around 1,600 enquiries a month from people considering pilot training, according to the group.
What the evidence reports
The Lufthansa Group's innovation pages say the academy launched George in October 2024 to answer questions about pilot training at any hour, and describe it as the group's first fully GPT-based bot.
Operational significance
An enquiry bot at a flight school sits well away from flight operations. It is still a live large language model answering the public on the group's behalf, where a wrong answer about training requirements has consequences for applicants.
Evidence still needed
The group has published no figures on how many enquiries George resolves, how often it hands over to staff or how its answers are checked. Other generative AI projects listed on the same pages, including NOTAM conversion and a pilot reference assistant, are at prototype or development stage.
Other records are shown separately, including non-AI technology and programmes at different stages of deployment. A register entry does not establish a measured AI benefit.
Optimized Passenger Recovery for rebooking and passenger notification at Lufthansa Group · Deployed · limited scope
When disruption hits, Lufthansa Group airlines use Optimized Passenger Recovery, an Amadeus tool that re-plans affected journeys. It checks seat availability, takes account of frequent-flyer status and follows the airlines' own business rules, such as which carriers to rebook passengers on and in what order.
That is all in a Lufthansa Group factsheet dated 1 October 2024. After rebooking, the factsheet says, the tool triggers the passenger notification, including the new boarding pass, for flights of the group's airlines, among them Lufthansa, SWISS, Austrian Airlines and Brussels Airlines.
Lufthansa's strongest AI case is in maintenance planning: a bounded task with a clear owner, human approval and figures the company has been willing to state, even though they are its own.
The next test is whether the planning gains reach the measures airline managers watch, such as delays caused by maintenance and aircraft availability. The group's longer list of generative AI projects is mostly at prototype stage and should be judged once those projects enter service.
Live data from the Aerostack register. Status reflects published evidence only. See Our approach for how records are graded.
Aerostack register · 3 tracked deployments
Aerostack tracks AI deployments across airports and airlines, with the published evidence behind each record. Photography is real aviation imagery. Record thumbnails may show the named airport or airline; other photographs do not.
