Which AI companies serve airlines and rail, and what they do
Six layers of what the companies sell, from the vehicle's own sensors up to the commercial plan.
The layers, derived from the companies
These layers were not copied from the manufacturing ladder. A boundary that most companies straddle is not a boundary; a layer with fewer than five companies (or three startups) was merged into its neighbour.
The manufacturing map keeps its own eight layers
- horizon
- seconds (driving) to months (the shop visit)
- who buys
- maintenance and engineering (TechOps, rolling-stock engineering), often together with the vehicle maker
- in
- engine data over its life cycle (Delta TechOps APEX); a rail vehicle's telemetry (Railnova); camera object detection on the locomotive (OTIV)
- out
- a component health alert (Alstom HealthHub, Hitachi Rail HMAX); a shop-visit and material plan 18 to 24 months ahead (Delta TechOps); the train driving itself between stations (OTIV, Alstom ATO)
- decision
- whether a component is about to fail and when to pull the vehicle; with ATO the train acts within its movement authority
- core stack
- analytics 2 · CV 1 · ML 1 (of 5)4 open tech roles at 2 companies: ML / data science 25%, OR 0%
data, standards and problem keys
- reads / writes
- reads engine, bogie and cab sensors and train telemetry · writes health alerts and shop-visit plans to the maintenance system, or traction and braking commands to the train (ATO)
- problem keys
- Predictive maintenance signals, Maintenance planning, Automatic train operation3 keys, 3 with a company
- horizon
- minutes (this turnaround) to a season (the stand plan)
- who buys
- airport operations, ground handlers, the operator's station and yard planning
- in
- camera feed of the aircraft stand (Assaia); turnaround events and IoT signals (ZestIoT); the day's arrivals and departures (Veovo, SITA Fixed Resource Manager); train units to park and service at a yard (NS HIP)
- out
- predicted off-block time and turnaround milestones (Assaia, ZestIoT); a stand and gate plan (Assaia StandManager, Veovo, INFORM GROUNDSTAR); ramp agents and equipment routed to the turnaround (Moonware HALO, BestTurn SpotWork); a yard parking and servicing plan (NS)
- decision
- which stand, gate, worker and yard track, and whether this turnaround will be late
- core stack
- workflow software 4 · OR 2 · CV 1 (of 9)22 open tech roles at 3 companies: ML / data science 5%, OR 0%
data, standards and problem keys
- reads / writes
- reads apron cameras, turnaround events and the day's arrivals and departures from the airport database · writes off-block predictions, stand and gate plans, ramp task assignments and yard plans
- standards
- A-CDM milestones, IATA AHMmixed: open shared standardno shared formatTurn-round milestones and ground messages have industry formats; predictions and camera-based turnaround data do not.
- problem keys
- Turnaround monitoring and prediction, Stand and gate assignment, Airport resource management, Shunting yard planning4 keys, 4 with a company
- horizon
- minutes to the end of the day
- who buys
- the operations control centre (OCC) duty manager, the rail dispatcher or traffic controller, flight dispatch
- in
- live train positions and the timetable (DB InfraGO ADA-PMB, Siemens TPS); airspace conditions up to eight hours ahead (Air Space Intelligence Flyways); the day's aircraft, crew, passengers and maintenance after a cancellation (Big Blue Analytics OCCam, StratosX, OpsLab)
- out
- a recommendation to the dispatcher every minute (DB InfraGO); a recovery timetable within minutes (NEC with Odakyu, China State Railway); ranked recovery plans for the duty manager (Big Blue, Lufthansa OPSD, Opfyx); a route and fuel plan for one flight (Flightkeys, Jeppesen); an answer from the official documents (Overwatch AI Wingman)
- decision
- the dispatcher or duty manager accepts a proposal; United's ConnectionSaver holds a flight for connecting passengers on its own
- core stack
- OR 7 · ML 3 · RAG 1 (of 12)1 open tech roles at 1 companies: ML / data science 0%, OR 0%
data, standards and problem keys
- reads / writes
- reads live positions, signalling or ATC state and the day's aircraft, crew and passengers from the operations control system · writes dispatcher proposals, recovery plans, one flight's route and fuel plan, and answers from the manuals
- standards
- ETCS and ATO in the CCS TSI, A-CDM, IATA delay codes in the AHM, EU flight time rulesmixed: open shared standarda big vendor's systemno shared formatTrain control and airport turn-round sharing are open standards, one backed by EU law. A recovery tool plugs into the airline's crew and operations system, the vendor's record. How advice reaches a driver or dispatcher has no standard that was found.
- problem keys
- Conflict detection and resolution, Train rescheduling, Driver advisory (energy-optimal speed), Arrival and departure sequencing, Disruption recovery, Day-of crew re-rostering, Flight planning, Delay prediction, Question answering over manuals and rules9 keys, 7 with a company
- horizon
- days to a month ahead (the monthly roster)
- who buys
- crew planning, fleet and rotation planning, resource planning inside the operator
- in
- the schedule, the fleet, crew legality rules and preferences (Jeppesen, Lufthansa Systems NetLine, SkAI Tech); timetable data and work rules of a Japanese railway (STRIDE, on GIRO's HASTUS engine)
- out
- pairings and monthly rosters (Motulus Aero, China Southern's own solver, STRIDE for Osaka Metro); fleet and tail assignment in the days before operations (KLM Pathfinder); rolling-stock rotations (DB Fleet Employment Optimization, EUR 74m a year); a daily forecast of reserve crew and open flying (Southwest)
- decision
- which named crew member and which tail or unit works which duty; planners accept and publish
- core stack
- OR 13 · analytics 1 · workflow software 1 (of 15)1 open tech roles at 2 companies: ML / data science 0%, OR 0%
data, standards and problem keys
- reads / writes
- reads the published schedule, the fleet and the crew rules and preferences from the crew and fleet systems · writes rotations, pairings, rosters and reserve forecasts back to them
- standards
- EU flight time rulesmixed: a big vendor's systemno shared formatCrew and operations systems are each vendor's own record, and crew rules are written as law and agreements with no data format found.
- problem keys
- Crew pairing, Crew rostering, Reserve crew and open-time forecasting, Crew self-service and bidding, Fleet assignment, Aircraft rotation and tail assignment, Rolling stock circulation7 keys, 6 with a company
- horizon
- a season to a year ahead (the annual timetable, the rolling seasonal plan)
- who buys
- the timetable office, the infrastructure manager, the air navigation network manager; in airlines the schedule planners
- in
- path orders from railway undertakings and construction works (RMCon RailSys at ÖBB Infrastruktur); the trains of any number of operators in one capacity model (SMA Viriato); an airline schedule to test across connections, rotations, crew and gates (Odysee AI)
- out
- the season timetable (GIRO HASTUS, Siemens TPS, SNCF Réseau DGEX, SBB Cargo's wagonload timetable); a capacity model under the EU Capacity Regulation (SMA); the Network Rolling Seasonal Plan over about 350 airlines and 43 states (EUROCONTROL Network Manager); harmonised capacity restrictions and capacity supply (RailNetEurope CPT)
- decision
- what runs where and when on shared infrastructure; the timetable office, infrastructure manager or network manager decides and publishes
- core stack
- OR 7 · data eng 3 · simulation 3 (of 14)
data, standards and problem keys
- reads / writes
- reads path orders, capacity requests and construction works from the operators and the infrastructure database · writes the season timetable, the capacity model and the network plan, published to every operator
- standards
- IATA SSIM, railML, WASG, Regulation 95/93, Directive 2012/34/EU, RNE PCS, Telematics TSI, NeTExopen shared standardThe most standardised layer on the map: schedules and timetables have open formats, and slot and path allocation is written in law and run by the allocator. Nobody on the map sells a what-if tool across the parties.
- problem keys
- Timetabling, Rail capacity allocation, Air traffic flow management, Airport slot coordination4 keys, 3 with a company
- horizon
- years ahead (the network) to the minute of sale (the price)
- who buys
- network planning, revenue management and pricing: the commercial side of the operator
- in
- bookings, demand forecasts and competitor fares (PROS, FLYR, Kambr); the current schedule and fleet to re-plan (Optifly, Sabre Mosaic, Amadeus SkyWORKS)
- out
- prices and seat allocation set in real time (Fetcherr); over 44 million seats re-priced a day (Flix); a long-term network plan and in-season schedule scenarios (Optifly, Amadeus); passengers who volunteer moved off a full flight to free seats to sell (Volantio, over USD 20m a year at Alaska Airlines); a utilisation forecast for day trains (ÖBB-Personenverkehr)
- decision
- which routes and frequencies to sell and at what price
- core stack
- ML 4 · OR 3 · analytics 2 (of 9)14 open tech roles at 1 companies: ML / data science 43%, OR 0%
data, standards and problem keys
- reads / writes
- reads bookings, demand forecasts and the schedule from the passenger service system · writes fares, seat availability and network scenarios back to it
- problem keys
- Network planning, Revenue management, Offer optimisation and dynamic pricing3 keys, 3 with a company
The problems the sector solves
A problem is a named task the sector solves with software: pairing crew, building a timetable, sequencing departures.The list has 31 problems; 27 of them have at least one of the 67 companies on this map, and 4 have none. A company counts under every problem it lists, primary first; it lists only what its own pages say it solves.
| problem | layer · horizon | classic method | companies on the map | who, by group | technical core most common | delivery models |
|---|---|---|---|---|---|---|
| Maintenance planning | AR1 | MIP, rules | 1 of 67 companies → | operator in-house unit: 1 of 1 | ML 1 · analytics 1 | in-house, not sold 1 |
| Automatic train operation | AR1 | control | 4 of 67 companies → | startup: 1 of 4 incumbent vendor: 1 of 4 operator in-house unit: 2 of 4 | control 3 · CV 1 · ML 1 | software + hardware 1 · in-house, not sold 2 |
| Predictive maintenance signals | AR1 | ML anomaly detection, digital twins | 4 of 67 companies → | startup: 1 of 4 incumbent vendor: 2 of 4 operator in-house unit: 1 of 4 | analytics 4 · ML 1 · control 1 | software + hardware 1 · in-house, not sold 1 |
| Airport resource management | AR2 | rules, MIP, simulation | 6 of 67 companies → | startup: 3 of 6 incumbent vendor: 3 of 6 | workflow software 4 · ML 2 · analytics 2 | software only 2 · software + hardware 1 |
| Shunting yard planning | AR2 | local search, MIP, evolutionary algorithms | 2 of 67 companies → | incumbent vendor: 1 of 2 operator in-house unit: 1 of 2 | OR 2 | in-house, not sold 1 |
| Stand and gate assignment | AR2 | rules, MIP, constraint programming | 2 of 67 companies → | startup: 2 of 2 | CV 1 · ML 1 · OR 1 | software only 2 |
| Turnaround monitoring and prediction | AR2 | computer vision, ML | 2 of 67 companies → | startup: 2 of 2 | CV 1 · ML 1 · analytics 1 | software only 1 · software + hardware 1 |
| Disruption recovery | AR3 | MIP, local search, simulated annealing, recently ML-guided | 9 of 67 companies → | startup: 3 of 9 incumbent vendor: 3 of 9 operator in-house unit: 2 of 9 acquired startup: 1 of 9 | OR 6 · ML 3 · workflow software 3 | software only 3 · in-house, not sold 2 |
| Day-of crew re-rostering | AR3 | heuristics, MIP, rules | 2 of 67 companies → | startup: 2 of 2 | ML 2 · OR 2 · RAG 1 | software only 1 |
| Train rescheduling | AR3 | MIP, simulation, reinforcement learning | 2 of 67 companies → | incumbent vendor: 1 of 2 operator in-house unit: 1 of 2 | ML 1 · OR 1 · control 1 | in-house, not sold 1 |
| Flight planning | AR3 | shortest path, trajectory optimisation | 3 of 67 companies → | startup: 2 of 3 incumbent vendor: 1 of 3 | OR 2 · simulation 2 · ML 1 | software only 1 |
| Delay prediction | AR3 | ML (gradient boosting, neural nets) | 1 of 67 companies → | operator in-house unit: 1 of 1 | ML 1 · data eng 1 | in-house, not sold 1 |
| Driver advisory (energy-optimal speed) | AR3 | optimal control, rules | - | - | - | |
| Conflict detection and resolution | AR3 | MIP, heuristics, reinforcement learning | 3 of 67 companies → | incumbent vendor: 1 of 3 operator in-house unit: 2 of 3 | ML 2 · OR 2 · simulation 1 | in-house, not sold 2 |
| Arrival and departure sequencing | AR3 | rules, MIP, A-CDM pre-departure sequencer | - | - | - | |
| Question answering over manuals and rules | AR3 | retrieval-augmented generation | 3 of 67 companies → | startup: 2 of 3 incumbent vendor: 1 of 3 | RAG 3 · OR 2 · data eng 1 | software only 2 · software + heavy services 1 |
| Fleet assignment | AR4 | MIP, network flow | 3 of 67 companies → | startup: 1 of 3 incumbent vendor: 1 of 3 operator in-house unit: 1 of 3 | OR 3 · agents 1 · workflow software 1 | software only 1 · in-house, not sold 1 |
| Aircraft rotation and tail assignment | AR4 | MIP, column generation | 4 of 67 companies → | startup: 2 of 4 incumbent vendor: 1 of 4 operator in-house unit: 1 of 4 | OR 4 · simulation 1 | software only 2 · in-house, not sold 1 |
| Rolling stock circulation | AR4 | hypergraph MIP, hierarchical column generation | 2 of 67 companies → | incumbent vendor: 1 of 2 operator in-house unit: 1 of 2 | OR 2 | in-house, not sold 1 |
| Crew pairing | AR4 | set partitioning with column generation (resource-constrained shortest path) | 7 of 67 companies → | startup: 3 of 7 incumbent vendor: 2 of 7 operator in-house unit: 2 of 7 | OR 7 · analytics 1 · simulation 1 | software only 2 · software + heavy services 1 · in-house, not sold 2 |
| Crew rostering | AR4 | MIP, heuristics, preferential bidding | 14 of 67 companies →namesOpfyx, China Southern Airlines (algorithm centre, crew optimisation system), DELMIA Quintiq (rail service, fleet and crew planning), IVU.suite (bus and rail), Jeppesen (crew, tail assignment, flight planning), Lufthansa Systems (NetLine and aiOCC), Wintelia (恒赢智航, iCREW and iFOC), Menlo79, Motulus Aero, Floating Fleet AI, SkAI Tech, GIRO (HASTUS, transit and rail scheduling), Optibus, Trapeze Group (rail timetable and crew planning) | startup: 5 of 14 incumbent vendor: 8 of 14 operator in-house unit: 1 of 14 | OR 13 · workflow software 4 · RAG 2 | software only 8 · software + heavy services 1 · in-house, not sold 1 |
| Reserve crew and open-time forecasting | AR4 | ML forecasting | 1 of 67 companies → | operator in-house unit: 1 of 1 | OR 1 · analytics 1 | in-house, not sold 1 |
| Crew self-service and bidding | AR4 | preference-based assignment | - | - | - | |
| Timetabling | AR5 | periodic event scheduling, MIP, constraint programming | 11 of 67 companies →namesGIRO (HASTUS, transit and rail scheduling), Indian Railways (Centre for Railway Information Systems, CRIS), Optibus, SBB Cargo (Operations Research and Analytics), Siemens Mobility Train Planning System (TPS suite), SNCF Réseau (DGEX timetabling research), Trapeze Group (rail timetable and crew planning), RMCon International (RailSys), SMA und Partner (Viriato), Odysee AI, Biarri Rail | startup: 1 of 11 incumbent vendor: 6 of 11 operator in-house unit: 3 of 11 acquired startup: 1 of 11 | OR 9 · ML 3 · simulation 3 | software only 4 · software + heavy services 2 · in-house, not sold 3 |
| Rail capacity allocation | AR5 | rules, negotiation, auction, MIP | 3 of 67 companies → | incumbent vendor: 3 of 3 | simulation 2 · workflow software 2 · OR 1 | software + heavy services 2 |
| Air traffic flow management | AR5 | rules, optimisation, simulation | 2 of 67 companies → | incumbent vendor: 2 of 2 | data eng 1 · simulation 1 | - |
| Airport slot coordination | AR5 | rules (Worldwide Airport Slot Guidelines), priority algorithms | - | - | - | |
| Network planning | AR6 | OR and demand models (MIP, simulation) | 5 of 67 companies → | startup: 1 of 5 incumbent vendor: 3 of 5 operator in-house unit: 1 of 5 | OR 5 · workflow software 2 · ML 1 | software only 2 · in-house, not sold 1 |
| Revenue management | AR6 | OR (EMSR, network optimisation) plus demand forecasting | 9 of 67 companies →namesAmadeus (airline revenue management, disruption and schedule tools), Flix (network planning, pricing and yield platform), FLYR, PROS (airline offer and revenue management), Sabre Mosaic (airline revenue and offer management), ÖBB-Personenverkehr (Operations Research team in yield management), Volantio, Fetcherr, Kambr | startup: 2 of 9 incumbent vendor: 4 of 9 operator in-house unit: 2 of 9 acquired startup: 1 of 9 | ML 4 · OR 3 · analytics 2 | software only 5 · in-house, not sold 2 |
| Offer optimisation and dynamic pricing | AR6 | ML demand and price models | 5 of 67 companies → | startup: 1 of 5 incumbent vendor: 3 of 5 operator in-house unit: 1 of 5 | ML 3 · OR 2 | software only 3 · in-house, not sold 1 |
| Operations data platform | flag | data engineering | 3 of 67 companies → | startup: 2 of 3 incumbent vendor: 1 of 3 | analytics 3 · ML 1 · data eng 1 | software only 1 · software + hardware 1 |
Click a count to filter the Companies tab to that problem; open "names" for the rows.
Where the standard lives
A decision's data format lives in one of three places: an open standard, a big vendor's system, or no shared format. Where an open standard exists, a new tool can plug in quickly. Where one vendor owns the record or the control interface, such as the airline's crew and operations system, that vendor tends to keep the margin. Where nothing exists, every operator is a custom integration project.
Terms used here (18), one line each
Systems, data and operations words
- PSS
- passenger service system: the airline's reservation, inventory and ticketing system.
- crew management system
- the system that holds crew rosters, qualifications and legality rules and assigns crew to flights or trains.
- OCC
- operations control centre: where an airline manages the day's flying and handles disruption.
- slot
- a permission to take off or land at a congested airport at a given time.
- slot coordinator
- the neutral body that allocates slots at a congested airport.
- infrastructure manager
- the company that owns and runs the track and signalling and allocates train paths to operators.
- path
- a train's reserved slot on the track, with times at each point.
- ETCS
- European Train Control System: the standard signalling system that supervises train speed in the cab.
- ATO
- automatic train operation: the train drives itself within the limits ETCS sets.
- TSI
- technical specification for interoperability: an EU regulation setting what a rail subsystem must do to run across borders.
- ATFM
- air traffic flow management: regulating how many flights enter a sector or airport per hour.
- A-CDM
- airport collaborative decision making: airport, airlines, ground handlers and air traffic control share departure times.
- crew pairing
- building the chains of flights, from base to base, that one crew will fly.
- rostering
- assigning crew to the pairings or duties for a month.
- tail assignment
- deciding which aircraft, by registration, flies which flight.
- rolling stock
- the trains and carriages of a railway.
- turn-round
- the ground work between an aircraft's arrival and its next departure.
- timetable
- the planned times of every train or flight, built months ahead.
IATA SSIM (airline schedules and slot messages), railML (timetable and track data for railways), CCS TSI (ETCS) (the EU's legal signalling standard), A-CDM (airports sharing departure information), WASG and Regulation 95/93 (airport slot rules)
for whoever builds on top Anyone can use the format. Here many exist because a law or a slot coordinator requires them, so the body that allocates owns the format, not a software vendor. iata.org eurocontrol.int
the airline's crew and operations system (systems such as AIMS and Netline, named on SITA's OCC Assistant Manager page as what new tools must integrate with)
for whoever builds on top A new planning tool plugs into the airline's existing crew and operations system.SITA's page says so for its own product.
working-time and rest rules for crews, written as text (AR4), what a sensor reading on an engine or bogie means (AR1), what-if runs over slots and train paths across several parties (AR5)
for whoever builds on top Each deployment starts by turning rules written as text into data by hand. Whoever gets a format adopted would own it, but only if competitors adopt it too.
Layer by layer
For each layer: which standards already exist, and in plain words where the agreement lives.
| layer | standards that already exist | where it lives |
|---|---|---|
| AR1 The vehicle: health signals and self-driving | mixed: open shared standarda big vendor's system The train's self-driving is under the EU's legal signalling standard.Health data sits in the makers' and maintenance providers' own platforms. | |
| AR2 The node: airport, station and yard | A-CDM milestones (airports, airlines and air traffic control sharing departure times), IATA AHM (the industry manual for ground handling, including delay codes) | mixed: open shared standardno shared format Turn-round milestones and ground messages have industry formats; predictions and camera-based turnaround data do not. |
| AR3 The control room: today's flights and trains | ETCS and ATO in the CCS TSI (the EU's legal signalling standard (ETCS), including automatic train operation), A-CDM (airports, airlines and air traffic control sharing departure times), IATA delay codes in the AHM (the industry manual for ground handling, including delay codes), EU flight time rules (the EU law on how long pilots and cabin crew may work) | mixed: open shared standarda big vendor's systemno shared format Train control and airport turn-round sharing are open standards, one backed by EU law. A recovery tool plugs into the airline's crew and operations system, the vendor's record. How advice reaches a driver or dispatcher has no standard that was found. |
| AR4 The operator's resources: fleet, rolling stock and crew | EU flight time rules (the EU law on how long pilots and cabin crew may work) | mixed: a big vendor's systemno shared format Crew and operations systems are each vendor's own record, and crew rules are written as law and agreements with no data format found. |
| AR5 The timetable and the network's capacity | IATA SSIM (the standard format for airline schedules and slot messages), railML (open file format for timetables and track data), WASG (the global rulebook for airport take-off and landing slots), Regulation 95/93 (the EU law that allocates airport slots), Directive 2012/34/EU (the EU law behind rail access and timetabling), RNE PCS (the shared tool for requesting train paths across borders), Telematics TSI (the EU's legal data-sharing rules for rail), NeTEx (European format for public transport timetables) | open shared standard The most standardised layer on the map: schedules and timetables have open formats, and slot and path allocation is written in law and run by the allocator. Nobody on the map sells a what-if tool across the parties. |
| AR6 The commercial plan: network, prices and seats | NDC (airlines' shared format for selling fares and extras), IATA SSIM (the standard format for airline schedules and slot messages) | mixed: open shared standarda big vendor's system Offers and schedules have open exchange formats (NDC, SSIM). |
Each standard links to a page it was checked on; hover for the body and the year. Every acronym is spelled out under "Terms used here" above.
What drives standardization
Three things push an industry onto a standard: a big buyer that will not buy without it, a law, or a group of companies in one industry agreeing among themselves. In airlines and rail the law and the allocating body carry most of it. Of the 3 cases below, the push came 1 by a standards body, 1 by big buyers' installed systems, 1 by regulation. These are 3 cases, so they illustrate and do not measure.
3 more cases
Europe's governments wrote one set of rules for who gets a take-off and landing slot at a crowded airport. The airline industry's own guidelines now cover the same ground worldwide.
evidence The regulation is still in force after amendment; IATA calls WASG the globally accepted framework.
who Council of the European Communities; airlines, airports and coordinators through IATA, ACI and WWACG
A new disruption-recovery tool is sold as something that plugs into the airline's existing crew and operations systems instead of replacing them.
evidence SITA's page says the product integrates with AIMS and Netline without replacing them.
who Crew and operations system vendors; SITA with Big Blue Analytics as the newcomer that integrates
Large European airports adopted one shared way of exchanging departure information without being forced by law, 35 of them fully so, according to EUROCONTROL.
evidence EUROCONTROL: fully implemented at 35 airports across Europe, on a voluntary footing as far as the page shows. No share of all airports is given.
who EUROCONTROL with airport operators, aircraft operators, ground handlers and air traffic control
What should be standardized next
- AR4 Crew working-time rules in a form software can read
The law says how long crews may work. It is written as text, and no data format for it was found.
Regulation (EU) No 83/2014 sets flight time limits for air operators from 18 February 2016.No machine-readable rule format was found.
- AR5 Capacity numbers coordinators can share for what-if planning
Requests and allocations of slots and train paths have standards.
WASG, Regulation 95/93, RNE PCS and the Telematics TSI cover requests and allocations.
This is the gap this map was built to show. 4 of 67 companies carry the between-organisations flag: their pages show them modelling parties outside one operator (DB InfraGO (ADA-PMB dispatching assistant), RMCon International (RailSys), SMA und Partner (Viriato) and R-NEST (EUROCONTROL Research Network Strategic Tool)). They stay on the layer of the service they sell: 1 on AR3 (The control room: today's flights and trains), 3 on AR5 (The timetable and the network's capacity).
- AR2 Delay causes counted the same way everywhere
Airlines have an industry manual for delay codes.
AHM 732 covers delay code procedures (47th edition, 2027).
- AR3 A standard way to give a driver or dispatcher advice
Train control has a legal standard.
The CCS TSI covers ETCS and ATO baseline 1. It does not cover advice; none was found.
- AR1 Shared meaning for engine and bogie signals
Who sits where
Each company appears once, on its main layer; tick the box to also see it, faded, on its secondary layers. The chip style says how large the company is and how public its customer record is, by the rules in the legend; a letter before the name says whether it is an incumbent vendor, an operator's in-house unit or an acquired startup; a dashed purple outline is the between-organisations flag.
4 of 67 companies carry the between-organisations flag: their pages show them modelling parties outside one operator (DB InfraGO (ADA-PMB dispatching assistant), RMCon International (RailSys), SMA und Partner (Viriato) and R-NEST (EUROCONTROL Research Network Strategic Tool)). They stay on the layer of the service they sell: 1 on AR3 (The control room: today's flights and trains), 3 on AR5 (The timetable and the network's capacity).
Layer × who closes the loop
Each dot is one company at its main layer. The vertical position says who acts on the output: the system itself at the top, a human reading a report at the bottom.
Companies by layer
| Alstom (HealthHub and automatic train operation) | incumbent vendor | rail | AR1 | Rolling-stock and signalling maker; HealthHub for predictive maintenance and Automatic Train Operation for driverless metro and regional trains. | Predictive maintenance signalsAutomatic train operation | analytics · controlwhyAlstom's digital mobility page lists HealthHub as maintenance intelligence and Automatic Train Operation as a train-control product. hiring no count: alstom.com/careers fetched (HTTP 200) but renders its job list in JavaScript; no counts could be read, and company-wide postings would not name HealthHub or ATO. sources alstom.com | Established | reads telemetry | 86,039 Wikipedia: 86,039 employees (March 2025) https://en.wikipedia.org/wiki/Alstom | alstom.com en.wikipedia.org | |
| Delta TechOps (APEX engine predictive maintenance) | operator in-house unit | airline | AR1 | Delta TechOps' APEX collects real-time data as an engine goes through its life cycle and uses it, with AI, to optimise engine material and shop visits; predictive material demand accuracy rose from 60 to over 90 percent, and shop visits are planned 18 to 24 months ahead. | Maintenance planningPredictive maintenance signals | ML · analyticswhy"Predictive analytics system powered by artificial intelligence" (Aviation Week); no model named. sources mroevents.aviationweek.com | detailbuyer Delta Air Lines itself (own operation) "Eight-digit cost savings" for the airline (Delta TechOps page). | Established | reads telemetry | ? | deltatechops.com mroevents.aviationweek.com leadiq.com |
| Hitachi Rail HMAX (digital asset management) | incumbent vendor | rail | AR1 | Rail systems maker; HMAX asset-management products for trains, signalling and infrastructure support predictive maintenance. | Predictive maintenance signals | analytics | Established | reads telemetry | 24,000 own site https://www.hitachirail.com/: 24K total employees | hitachirail.com en.wikipedia.org | |
| OTIV | startup | rail | AR1 | Rail automation stack: shunting camera system, driver assistance, remote supervision centre and full autonomous driving, built on camera object detection. | Automatic train operation | CV · edge / IIoT · controlwhyCamera-based perception for shunting, driver assistance and autonomous driving (OTIV.ONE to .FOUR, otiv.ai); machine learning engineer, hardware engineer and robotics software roles open. hiring 4 open tech roles: ML 1, software 1, embedded / edge 1, other 1 apply.workable.com. Other = Internship Engineering.) sources otiv.ai apply.workable.com | detailbuyer industrial companies, railway undertakings, infrastructure managers and OEMs (site wording; a customer type, not a job role) | Customer track record | controls the aircraft, train or sequence | ? | otiv.ai tech.eu |
| Railnova | startup | rail | AR1 | SaaS that connects rail vehicles' telemetry, diagnoses faults and monitors fleets so operators can maintain and run them. | Predictive maintenance signalsOperations data platform | edge / IIoT · data eng · analyticswhyTelemetry hardware plus a platform that applies 1,400+ rules to raw vehicle signals for fault diagnosis and fleet monitoring (railnova.eu/en/about-us). hiring 0 open tech roles jobs.railnova.eu (The careers page is live and its Recruitee offers API returned an empty list, so no open roles today.) sources railnova.eu | detailAbout page says hardware designed in-house. | Customer track record | reads telemetry | ? | railnova.eu |
| INFORM GROUNDSTAR (airport ground operations) | incumbent vendor | air traffic and airports | AR2 | Operations-research software company; GROUNDSTAR is its airport and ground-handling resource management product. | Airport resource management | OR · MLwhyINFORM's German careers page: software for process optimisation 'based on artificial intelligence and operations research' (company level, not GROUNDSTAR specific). hiring 17 open tech roles: software 9, solutions / deployment 5, other 3 career.inform-software.com. Left out: 2 apprenticeship or dual-study, 4 sales, office manager, 2 controllers.) | Established | reads ops, crew or timetable systems | 968 German Wikipedia: 968 employees (2023) https://de.wikipedia.org/wiki/Inform_(Unternehmen); company-wide, not the aviation unit | inform-software.com de.wikipedia.org | |
| NS (shunting yard planning, HIP) | operator in-house unit | rail | AR2 | NS deployed the Hybrid Integrated Planning Method (HIP) for parking and servicing planning at shunting yards, developed with Utrecht University; live at 3 yards of 32. | Shunting yard planning | ORwhyHybrid Integrated Planning Method for parking and servicing at shunting yards, developed with Utrecht University. sources uu.nl | detailbuyer NS itself (own operation) | Established | reads ops, crew or timetable systems | ? | uu.nl |
| SITA (Airport Management, Fixed Resource Manager) | incumbent vendor | air traffic and airports | AR2 | Air transport IT provider owned by about 400 industry members; its Airport Management suite includes Fixed Resource Manager, "a powerful scheduling, capacity planning, and resource management solution" with scenario-based planning and day-of-operations planning. | Airport resource management | workflow softwarewhyProduct page names scheduling, capacity planning and scenario-based planning; no method named. sources sita.aero | Established | reads ops, crew or timetable systems | ? ICAO postal-history page: "some 4,500 employees" (undated); neither used as the number | sita.aero icao.int en.wikipedia.org | |
| Veovo (airport operations platform) | incumbent vendor | air traffic and airports | AR2 | Airport operations software (AODB, resource management, A-CDM, passenger flow forecasting) used by over 140 airports; resource management plans stands, gates and other resources to maximise capacity and minimise delays. | Airport resource managementOperations data platform | workflow software · analytics | detailDelivery model inferred from the product list only. | Established | reads ops, crew or timetable systems | ? | veovo.com |
| Cosmos | startup | airline | AR2 | Platform that monitors service level agreements between airlines and service providers with machine learning, finds causes of delays from inefficient processes and enables real-time communication. | Operations data platform | ML · analyticswhyaeroTELEGRAPH announcement: ML-based monitoring of service level agreements between airlines and service providers, finding causes of delays. sources aerotelegraph.com | - | Early | reads ops, crew or timetable systems | ? | aerotelegraph.com |
| Moonware | startup | air traffic and airports | AR2 | HALO, an airside operating system that schedules and routes ramp agents and equipment for aircraft turnarounds and coordinates ground tasks. | Airport resource management | workflow software · edge / IIoT · MLwhyHALO assigns and reassigns ground resources, geolocates GSE, personnel and aircraft, and predicts what is next (moonware.com home); careers list one ML engineer role and Android and infrastructure roles. hiring 5 open tech roles: ML 1, software 4 moonware.com (Tech Lead, Senior SWE Machine Learning, Senior SWE Infrastructure, SWE Android, Senior SWE Full-stack; all Los Angeles.) sources moonware.com moonware.com | - | Early | reads ops, crew or timetable systems | ? | moonware.com newswire.com |
| Assaia | startup | air traffic and airports | AR2 | Computer vision on apron cameras that gives airports, airlines and ground handlers real-time visibility of aircraft turnaround operations. | Turnaround monitoring and predictionStand and gate assignment | CV · MLwhyHomepage: 'AI and computer vision' on existing apron cameras that timestamp each part of an aircraft's time on the ground; ML for predictive off-block time named in an Aeroporti di Roma testimonial. hiring 0 open tech roles assaia.com. The company states it is 100% remote and lists spontaneous applications for DE, ME, RS, ES, US.) sources assaia.com assaia.com | detailbuyer airport and airline operations; testimonial signers on the homepage are the Chief Aviation Officer of Aeroporti di Roma and the SVP Aviation of Flughafen Berlin Brandenburg Delivery model: homepage says the product runs on existing apron cameras. | reads telemetry | ? | assaia.com assaia.com arabfounders.net | |
| BestTurn (Beoseuteon) | startup | airline | AR2 | Operates aircraft turnaround ground services at Incheon (cleaning, wheelchair, escort) and sells SpotWork, a platform that matches credentialed on-demand ground workers to urgent airport tasks using certification, location and fatigue data. | Airport resource management | workflow softwarewhySpotWork is described as matching credentialed on-demand ground workers to urgent airport tasks (Ubergizmo article); no algorithm or ML detail found. sources ubergizmo.com bestturnaround.com | -detailAlso operates ground services itself. | reads ops, crew or timetable systems | ? | ubergizmo.com bestturnaround.com bestturnaround.com | |
| ZestIoT | startup | air traffic and airports | AR2 | AI and IoT platform for airlines, airports and ground handlers that digitises the aircraft turnaround, predicts off-block delays with alerts, and plans and dispatches ground resources in real time. | Turnaround monitoring and predictionAirport resource management | edge / IIoT · analyticswhyOwn pages name IoT tracking, a turnaround digital twin, dashboards and AI prediction; no method named. sources zestiot.com | detailDelivery model: IoT tracking of assets plus software and a mobile app, per the airlines page. | reads telemetry | ? | zestiot.com zestiot.com | |
| China State Railway Group (intelligent high-speed rail dispatching, CTC research) | operator in-house unit | rail | AR3AR1 | A China State Railway Group research project produced an intelligent centralised traffic control system that adjusts the train running plan automatically after large delays with a multi-objective model, automates route setting and passes the plan to ATO; applied on the Beijing-Zhangjiakou and Beijing-Shenyang high-speed lines. | Train reschedulingAutomatic train operation | OR · controlwhyPage names a multi-objective real-time adjustment model with intelligent algorithms, and route and ATO control functions. sources wap.china-railway.com.cn | detailbuyer China State Railway Group itself (own operation) | Established | reads ops, crew or timetable systems | ? | wap.china-railway.com.cn |
| DB InfraGO (ADA-PMB dispatching assistant) | operator in-house unit | rail | AR3 | DB InfraGO's AI dispatching assistant combines mathematical optimisation, heuristics and machine learning to give train dispatchers real-time recommendations; piloted on the Berlin S-Bahn in 2020, now active in four regions, advice only, the dispatcher decides. | Conflict detection and resolution | OR · MLwhy"Combines mathematical optimisation, heuristics and machine learning" (Railway Gazette). sources railwaygazette.com | detailbuyer DB InfraGO itself (own operation) | Established | reads ops, crew or timetable systems | ? | railwaygazette.com |
| Deutsche Bahn (Digitale Schiene Deutschland, Capacity and Traffic Management System) | operator in-house unit | rail | AR3AR1 | DB programme building an AI capacity and traffic management system for dispatching, trained by deep reinforcement learning in railway simulation, first targeted at ATO on the Stuttgart Digital Node. | Conflict detection and resolutionAutomatic train operation | ML · simulationwhyDSD AI page: AI trained by trial and error (deep reinforcement learning) in railway simulation. hiring no count: db.jobs fetched (redirects to db.jobs/de-de, HTTP 200) but job list is JavaScript; no postings naming the programme read. sources digitale-schiene-deutschland.de | Established | reads ops, crew or timetable systems | ? | digitale-schiene-deutschland.de digitale-schiene-deutschland.de | |
| IBS Software (iFlight) | incumbent vendor | airline | AR3 | Travel software company; iFlight is described as AI-driven flight, crew and disruption management for airlines. | Disruption recovery | - | - | Established | reads ops, crew or timetable systems | 5,000 own site https://www.ibsplc.com/about-us: '5,000+ professionals' | ibsplc.com ibsplc.com |
| Lufthansa Group (Operations Decision Support Suite, OPSD) | operator in-house unit | airline | AR3 | Cloud operational planning tool that joins data from aircraft rotation, passenger, crew and technical fleet systems and uses AI to suggest scenarios to the operations control centres of the group's airlines; first at SWISS, then rolled out across the group. | Disruption recovery | ML · data eng | detailbuyer Lufthansa Group itself (own operation) | Established | reads ops, crew or timetable systems | ? | lufthansagroup.com cloud.google.com |
| NEC (recovery timetable AI for rail, deep reinforcement learning) | incumbent vendor | rail | AR3 | NEC built a system that drafts recovery timetables after rail disruptions with deep reinforcement learning trained on a railway digital twin. NEC names crew planning and regular timetables as next steps. | Train rescheduling | ML · simulation | -detailResearch system at verification stage; no sales model stated. | Established | reads ops, crew or timetable systems | ? | jpn.nec.com nikkei.com iotnews.jp |
| United Airlines (ConnectionSaver) | operator in-house unit | airline | AR3 | United's ConnectionSaver "automatically identifies departing flights that can be held for connecting customers", weighing gate-to-gate walking time against the effect of a hold on other flights and customers. | Disruption recovery | - | detailbuyer United Airlines itself (own operation) | Established | reads ops, crew or timetable systems | ? | prnewswire.com |
| Air Space Intelligence | startup | airline | AR3 | Flyways predicts how airspace conditions will evolve up to eight hours ahead and recommends more efficient routes to airline dispatchers; the company now also works with the FAA on air traffic control and with defence and grid operators. | Flight planning | simulation · MLwhyHomepage: AI engines that model physical reality and simulate what happens next; TNMT: predicts airspace conditions. No method detail read. sources airspace-intelligence.com | detailDelivery model inferred from "software and AI contract" wording on the homepage. | Customer track record | reads ops, crew or timetable systems | - | airspace-intelligence.com tnmt.com |
| Overwatch AI | startup | airline | AR3 | Wingman, an AI assistant that answers pilots' and operations staff's questions in natural language from official documents and live operational data. | Question answering over manuals and rules | RAG · data engwhyWingman answers questions in natural language from manuals, maintenance records, flight data and weather and NOTAM APIs, with data ingestion and continuous evaluations (overwatch-ai.com). hiring no count: No careers or jobs link on the site (nav: Solution, Use Cases, Benefits, About, FAQ, Blog, Book a demo). sources overwatch-ai.com | detailSite presents a platform deployed to airline crews, ground ops and OCC. | Early | reads ops, crew or timetable systems | ? | startbase.de techstars.com pulse2.com |
| StratosX | startup | airline | AR3 | X360 recovery platform that detects emerging disruption and proposes recovery options for aircraft routing, crew deployment and passenger rebooking. | Disruption recoveryDay-of crew re-rostering | OR · ML · RAGwhyX360 solves aircraft, crew and passengers as one problem and ranks recovery plans by cost, predicts disruption risk up to seven days ahead, and Wingman answers from manuals and live data (stratosx.ai). hiring no count: Company page has an 'Open roles' button that goes to a contact form; no listings. sources stratosx.ai stratosx.ai | -detailbuyer airline operations control centre (site: 'What's your OCC up against?') Site offers a free 30-day trial and phased implementation with the team working alongside the operator; delivery model not set. | Early | reads ops, crew or timetable systems | 10 press: about 10 people, https://betakit.com/how-air-transat-hopes-stratosxs-ai-will-help-it-fight-flight-disruptions/ | betakit.com |
| Big Blue Analytics | acquired startup | airline | AR3 | Built OCCam (OCC Assistant Manager), an optimisation engine that re-plans aircraft, crew, passengers and maintenance together after a disruption and ranks recovery plans for the duty manager. | Disruption recovery | ORwhySITA's OCCam page: an optimisation engine that re-plans aircraft, crew, passengers and maintenance together after a disruption and ranks plans. hiring no count: No careers page: bigblueanalytics.com redirects to SITA (lane pass); the company was acquired by SITA, announced 2 June 2026. sources sita.aero | detailbuyer airline operations control centre duty manager (SITA page: ranks recovery plans for the duty manager) | reads ops, crew or timetable systems | ? | sita.aero altexsoft.com eglobaltravelmedia.com.au | |
| Flightkeys | startup | airline | AR3 | Flight planning and trajectory optimisation software and services for airlines (Flightkeys 5D), an Austrian-based flight planning service provider. | Flight planning | ORwhyHomepage: trajectory optimisation, flight path optimisation, flight planning and in-flight optimisation. hiring 1 open tech role: data eng 1 flightkeys.aero. sources flightkeys.aero flightkeys.aero | - | reads ops, crew or timetable systems | ? | flightkeys.aero flightkeys.aero | |
| Opfyx | startup | air traffic and airports | AR3AR2AR4 | Cloud operations platform for airlines, airports and ground handlers with a disruption solver giving explainable recovery options, allocation of aircraft, stands, staff and equipment, planning and rostering, and an AI hub for SOPs and manuals. | Disruption recoveryStand and gate assignmentCrew rosteringQuestion answering over manuals and rules | OR · workflow software · RAGwhyDisruption solver and allocation modules (optimisation), task and communication workflow, AI-powered docs library. sources opfyx.com | reads ops, crew or timetable systems | ? | opfyx.com opfyx.com | ||
| OpsLab | startup | airline | AR3 | Crew tracking and disruption recovery software that bolts onto an airline's crew management system and resolves mass cancellations with operations research solvers, plus shift scheduling for ground operations and dispatchers; also sells to air force squadrons. | Disruption recoveryDay-of crew re-rostering | OR · ML | reads ops, crew or timetable systems | ? | opslab.com opslab.com | ||
| Air France-KLM (KLM Pathfinder schedule optimisation) | operator in-house unit | airline | AR4 | Schedule optimisation tool for fleet and tail assignment developed by KLM and Boston Consulting Group, in a partnership since 2018, extended in 2021 and 2024. | Fleet assignmentAircraft rotation and tail assignment | ORwhyBCG page describes Pathfinder as a schedule optimisation tool for fleet and tail assignment. sources bcg.com | Established | reads ops, crew or timetable systems | ? | bcg.com | |
| American Airlines (Operations Research and Advanced Analytics) | operator in-house unit | airline | AR4 | American's OR and advanced analytics team, successor to American Airlines Decision Technologies, which built the TRIP crew-pairing system in the 1980s and 1990s. | Crew pairing | OR · analyticswhy1991 Interfaces paper by American Airlines Decision Technologies on the TRIP crew-pairing system (optimisation); the current team name includes 'Advanced Analytics' (Georgia Tech capstone page). | detailNo price applies. The 1991 paper says TRIP was adopted by 10 airlines and one railroad, which is an outside-use fact, not a price. | Established | reads ops, crew or timetable systems | ? | econpapers.repec.org capstone.isye.gatech.edu |
| China Southern Airlines (algorithm centre, crew optimisation system) | operator in-house unit | airline | AR4 | China Southern's in-house algorithm centre built its own optimisation algorithms for crew pairing and rostering so that the airline could run on a domestic solver instead of a foreign one, with plan costs reported on par with the international level. | Crew pairingCrew rostering | ORwhyPage names linear and mixed-integer programming and in-house optimisation algorithms. sources zn.caac.gov.cn | detailbuyer China Southern Airlines itself (own operation) | Established | reads ops, crew or timetable systems | 30 CAAC Central and Southern Regional Administration page, 2024-07-19: a team of "more than 30 people" (30多人), built around Dr Zhao Mingyu's studio and the algorithm centre with user departments; recorded as 30 | zn.caac.gov.cn |
| DELMIA Quintiq (rail service, fleet and crew planning) | incumbent vendor | rail | AR4 | Planning and optimisation suite for rail freight, passenger and infrastructure operators covering service offering, fleet circulation and crew duty plans. | Crew rosteringRolling stock circulation | ORwhyDassault product page: rail service, fleet and crew planning, described as planning and optimisation with integrated multi-discipline planning. hiring no count: 3ds.com/careers fetched but company-wide; no postings naming Quintiq rail planning read. sources 3ds.com | Established | reads ops, crew or timetable systems | 25,000 parent Dassault Systèmes: 25,000 employees (2025), Wikipedia https://en.wikipedia.org/wiki/Dassault_Syst%C3%A8mes; division headcount not public | 3ds.com casestudies.com en.wikipedia.org | |
| Deutsche Bahn (Fleet Employment Optimization, FEO) | operator in-house unit | rail | AR4 | DB's rolling stock rotation planning system, built with FU Berlin, Zuse Institute Berlin and LBW Optimization using hierarchical column generation; DB attributes EUR 74m annual savings to it. | Rolling stock circulation | ORwhy2021 Interfaces paper abstract: rolling stock rotation planning by hierarchical column generation. hiring no count: db.jobs job list is JavaScript; no postings naming FEO read. sources econpapers.repec.org | Established | reads ops, crew or timetable systems | ? 18 paper authors from DB divisions and partners, not a unit headcount | econpapers.repec.org | |
| IVU.suite (bus and rail) | incumbent vendor | airlines and rail | AR4 | Integrated IT for bus and rail operators: service planning, resource and duty planning, dispatching, fleet management. | Crew rostering | ORwhyIVU optimisation page: duty and vehicle working optimisation algorithms developed with Zuse Institute Berlin; scenario calculation for tenders. hiring 1 open tech role: solutions / deployment 1 ivu.com. Counted: Project Engineers Rail Software Implementation. Left out: business development, sales and key account manager, project manager. sources ivu.com | detailbuyer dispatch managers and planners at bus and rail operators (optimisation page wording) | Established | reads ops, crew or timetable systems | 1,069 German Wikipedia: 1,069 employees, annual report 2024 cited https://de.wikipedia.org/wiki/IVU_Traffic_Technologies | ivu.com de.wikipedia.org |
| Jeppesen (crew, tail assignment, flight planning) | incumbent vendor | airline | AR4AR3 | Aviation software and data company with crew pairing, crew rostering, tail assignment, ops control and flight planning products. | Crew pairingCrew rosteringAircraft rotation and tail assignmentFlight planning | OR · simulationwhyHomepage lists Crew Pairing, Crew Rostering, Tail Assignment, Aircraft Routing and a Total Airspace and Airport Modeler; no method named. hiring no count: Careers redirect to jeppesen.referrals.selectminds.com, which renders no job list without JavaScript. sources ww2.jeppesen.com | Established | reads ops, crew or timetable systems | 3,200 Wikipedia infobox: 3,200 employees https://en.wikipedia.org/wiki/Jeppesen | ww2.jeppesen.com en.wikipedia.org thomabravo.com | |
| Lufthansa Systems (NetLine and aiOCC) | incumbent vendor | airline | AR4AR6AR3 | Lufthansa Group's IT arm; NetLine network planning, fleet assignment, crew pairing and rostering suite, and aiOCC disruption management. | Crew pairingCrew rosteringNetwork planningFleet assignmentDisruption recovery | OR · workflow softwarewhysources lhsystems.com joc.com | - | Established | reads ops, crew or timetable systems | 2,800 Wikipedia: about 2,800 employees (February 2023) https://en.wikipedia.org/wiki/Lufthansa_Systems | lhsystems.com en.wikipedia.org joc.com |
| Southwest Airlines (Advanced Analytics and Optimization Solutions) | operator in-house unit | airline | AR4 | Team that built a daily forecast of reserve crew availability and open flying to guide crew scheduling premiums, launched August 2017. | Reserve crew and open-time forecasting | analytics · ORwhyDaily forecast of reserve crew availability and open flying; unit name includes optimization. sources lebow.drexel.edu | detailbuyer Southwest crew scheduling (own operation) | Established | reads ops, crew or timetable systems | ? | lebow.drexel.edu |
| Wintelia (恒赢智航, iCREW and iFOC) | incumbent vendor | airline | AR4AR3 | Chinese aviation IT vendor (Beijing Wintelia Technology) selling airline operations control (iFOC 3.0), intelligent crew rostering (iCREW 3.0) and LLM assistants for dispatch release and customer service to Chinese airlines and airports. | Crew rosteringQuestion answering over manuals and rules | OR · RAGwhyCrew rostering optimisation product and large-model assistants, per the CAAC News report. sources caacnews.com.cn | detailDelivery model: homepage lists consulting, testing, products and HR outsourcing services alongside software. | Established | reads ops, crew or timetable systems | 700 wintelia.com homepage: "700+ professional technical service team" | wintelia.com caacnews.com.cn |
| Menlo79 | startup | rail | AR4 | Cloud workforce management software (WILSON) for rail and logistics operators, for staffing and qualification management beyond what shift-planning tools and Excel handle. | Crew rostering | workflow softwarewhyWILSON is cloud workforce, staffing and qualification management for rail (rendered menlo79.com page: products WILSON and CLARA AI & Automation). hiring 0 open tech roles menlo79-gmbh.jobs.personio.de. sources menlo79.com | Customer track record | reads ops, crew or timetable systems | 35 press: https://invest-in.berlin/n/rail-tech-firm-menlo79-lands-major-investment-for-wilson-platform/ ('35-person company', Jan 2025) | invest-in.berlin railwaygazette.com | |
| Motulus Aero | startup | airline | AR4 | Airline crew scheduling and optimisation software: integrated crew solver, pairing optimiser, tail assignment and network optimisation, plus a crew web app (Crewpal) for roster input. | Crew pairingCrew rosteringAircraft rotation and tail assignment | ORwhyProprietary optimisation algorithms and solvers for crew scheduling, pairings, tail assignment and network optimisation (motulus.aero home and about pages). sources motulus.aero motulus.aero | detailSite lists products plus an optimisation-as-a-service offer for custom solutions. | Early | reads ops, crew or timetable systems | ? | motulus.aero theflyingengineer.com |
| STRIDE | startup | rail | AR4 | Introduces and operates an optimisation-based train crew duty scheduling system for Japanese railways, built on GIRO's HASTUS engine, and ran the same job as a service for Osaka Metro from 2024. | Crew pairing | ORwhyOptimisation-based train crew duty scheduling system built on GIRO's HASTUS engine. hiring no count: No careers page: stride-corp.com has only Home, company overview and news. sources prtimes.jp | detailbuyer railway crew planning (Osaka Metro is the named client) | Early | reads ops, crew or timetable systems | ? | prtimes.jp stride-corp.com |
| Floating Fleet AI | startup | airline | AR4 | Scheduling, crew assignment and crew rostering software for private aviation operators, expanding in 2026 to airlines and corporate flight departments; reduces deadhead flying. | Aircraft rotation and tail assignmentCrew rostering | OR | reads ops, crew or timetable systems | ? | floatingfleet.ai ainonline.com | ||
| SkAI Tech | startup | airline | AR4 | SaaS crew rostering, pairing optimisation and flight assignment for scheduled airlines, ACMI and business aviation operators, plugged into the operator's existing flight management system. | Crew rosteringCrew pairing | ORwhyHomepage describes an optimisation engine for rostering and pairing under FTL rules. sources skai-tech.fr | reads ops, crew or timetable systems | ? | skai-tech.fr | ||
| GIRO (HASTUS, transit and rail scheduling) | incumbent vendor | rail | AR5AR4 | Montreal scheduling software firm that began as a university operations research project; HASTUS covers timetabling, crew scheduling and rostering for public transport, with HASTUS-Rail applications since 2006. Bus-first; Long Island Rail Road has run HASTUS for scheduling since 2006. | TimetablingCrew rostering | OR · workflow softwarewhy"Optimisation algorithms at its core" (GIRO talk); MTA describes scheduling optimisation. sources rotterdam2015.caspt.org mta.info | detailMTA lists a noncompetitive HASTUS contract of USD 8,023,546 for LIRR, expiring March 2027 (not entered as a deal size; it bundles services). | Established | reads ops, crew or timetable systems | - | rotterdam2015.caspt.org mta.info leadiq.com |
| Indian Railways (Centre for Railway Information Systems, CRIS) | operator in-house unit | rail | AR5AR3 | The Ministry of Railways' own IT organisation, which builds Indian Railways' control office automation and crew management systems and states that it uses AI and ML for train timetable optimisation, train delay prediction and predictive maintenance. | TimetablingDelay prediction | data eng · MLwhyHomepage names AI, ML, big data and BI among the technologies used for timetable optimisation and delay prediction; no method detail. sources cris.org.in | detailbuyer Indian Railways itself (own operation) | Established | reads ops, crew or timetable systems | ? | cris.org.in |
| Optibus | incumbent vendor | rail | AR5AR4 | Cloud platform for public transport planning, scheduling, rostering and operations, "powered by AI and optimization algorithms"; bus-first, with light rail, trams and subways among the trips it plans. | TimetablingCrew rostering | OR · MLwhyIVC description: "powered by artificial intelligence (AI) and optimization algorithms". sources ivc-online.com | Established | reads ops, crew or timetable systems | - | ivc-online.com cbinsights.com cbinsights.com | |
| SBB Cargo (Operations Research and Analytics) | operator in-house unit | rail | AR5 | SBB Cargo's OR and analytics team, led by Carsten Moldenhauer, partners with EPFL on mathematical methods to draft wagonload timetables automatically (2020 to 2024, Innosuisse-funded). | Timetabling | OR | detailbuyer SBB Cargo itself (own operation) | Established | reads ops, crew or timetable systems | ? | epfl.ch |
| Siemens Mobility Train Planning System (TPS suite) | incumbent vendor | rail | AR5AR3AR2 | Siemens Mobility's TPS suite for infrastructure managers and railway operators: timetable creation, real-time dispatching with network-wide conflict detection and automatic resolution, track-work slot planning and yard optimisation. | TimetablingConflict detection and resolutionShunting yard planning | ORwhyPage describes optimisation of timetables, conflict resolution and yard planning; no solver named. sources mobility.siemens.com | Established | reads ops, crew or timetable systems | ? | mobility.siemens.com | |
| SNCF Réseau (DGEX timetabling research) | operator in-house unit | rail | AR5 | SNCF Réseau staff in the DGEX unit co-author a heuristic for periodic timetabling and track assignment in the capacity structuring phase of yearly timetables. | Timetabling | ORwhyHeuristic for periodic timetabling and track assignment co-authored by DGEX staff. sources tristanconference.org | detailbuyer SNCF Réseau itself (own operation) | Established | reads ops, crew or timetable systems | ? | tristanconference.org |
| Trapeze Group (rail timetable and crew planning) | incumbent vendor | rail | AR5AR4 | Transit software group whose rail products (TrainPlan, RailPlan, Crew Optimiser, ResourcePlan, from the 2014 purchase of Funkwerk Information Technologies York) build timetables, simulate railway operations, plan fleets and optimise crew allocation, from long-term planning to the day. | TimetablingCrew rostering | OR · simulationwhyProduct text: crew allocation optimisation and railway operations simulation. sources trapezegroup.com finance.yahoo.com | Established | reads ops, crew or timetable systems | ? | trapezegroup.com finance.yahoo.com trapezegroup.com.my | |
| RMCon International (RailSys) | incumbent vendor | rail | AR5 | Hannover rail consultancy and maker of the RailSys suite, in which infrastructure managers take path orders from railway operators, build and simulate the timetable, coordinate construction works and distribute the result; ÖBB Infrastruktur plans its timetable in it from 2025. | Rail capacity allocationTimetabling | simulation · ORwhySuite page names timetable and operations simulation, capacity optimisation and a digital twin of the infrastructure. sources rmcon-int.de | detailDelivery model: software suite plus consulting, per the company page. | Customer track record | reads ops, crew or timetable systems | - | rmcon-int.de rmcon-int.de rmcon-int.de |
| SMA und Partner (Viriato) | incumbent vendor | rail | AR5 | Swiss railway planning consultancy and maker of Viriato, a timetabling system in which infrastructure managers model network capacity with the trains of any number of operators and construction works, compare variants and publish capacity models under the EU Capacity Regulation. | Rail capacity allocationTimetabling | simulation · workflow softwarewhyPages describe variant comparison, robustness assessment and running-time calculation; no solver or ML named. sources sma-partner.com sma-partner.com | detailDelivery model: consulting and software sold together, per the homepage. | Customer track record | reads ops, crew or timetable systems | 85 sma-partner.com homepage: "employs approximately 85 members of staff" | sma-partner.com sma-partner.com sma-partner.com |
| Odysee AI | startup | airline | AR5 | Schedule and revenue intelligence platform that evaluates an airline schedule across connections, rotations, crew, gates and revenue and compresses scenario analysis from days to minutes. | Timetabling | ML · analytics · ORwhyRevenue and misconnection prediction from a model trained on the airline's data (100+ signals), connectivity analytics, and a schedule optimizer in early access (odysee.ai). sources odysee.ai | detailbuyer airline schedule planners and schedulers (site wording) Site says it sits alongside the existing scheduling system. | Early | reads ops, crew or timetable systems | ? | odysee.ai tnmt.com |
| Biarri Rail | acquired startup | rail | AR5 | Australian provider of planning and scheduling software for freight railroads, now Navis's Freight Rail Planning and Scheduling product. | Timetabling | ORwhysources accel-kkr.com | reads ops, crew or timetable systems | ? | accel-kkr.com | ||
| EUROCONTROL Network Manager | incumbent vendor | air traffic and airports | AR5 | The public body that runs European air traffic flow management and the Network Rolling Seasonal Plan, consolidating data from about 350 airlines, 68 area control centres, 55 airports and 43 states. | Air traffic flow management | data engwhyPage: consolidates data from airlines, ACCs, airports and states into a rolling plan; no method is stated. hiring no count: eurocontrol.int/careers fetched (HTTP 200), company-wide, no Network Manager postings read. sources eurocontrol.int | - | reads ops, crew or timetable systems | ? | eurocontrol.int | |
| R-NEST (EUROCONTROL Research Network Strategic Tool) | incumbent vendor | air traffic and airports | AR5 | A EUROCONTROL network simulation tool that runs ECAC-wide scenarios over airspace, ATC centres, airports and airline behaviour to assess where ATM concepts and capacity measures should be deployed; used by EUROCONTROL and research partners, not sold. | Air traffic flow management | simulationwhyNetwork simulation over airspace, ATC centres, airports and airline behaviour. hiring no count: Public research tool, no unit-specific postings searched. sources eurocontrol.int | - | reads ops, crew or timetable systems | ? | eurocontrol.int eurocontrol.int | |
| RailNetEurope Capacity Planning Tool (CPT) | incumbent vendor | rail | AR5 | RNE's Capacity Planning Tool, complete since 19 October 2026, brings temporary capacity restrictions, capacity models and capacity supply into one shared planning environment where infrastructure managers harmonise capacity and applicants announce future traffic needs. | Rail capacity allocation | data eng · workflow softwarewhyShared planning database with publication, coordination and a Telematics-TSI message interface, per the RNE page. sources rne.eu | -detailProvided by an association to its members; no price applies or is stated. | reads ops, crew or timetable systems | ? | rne.eu | |
| Amadeus (airline revenue management, disruption and schedule tools) | incumbent vendor | airline | AR6AR3 | Travel technology group whose airline products include Network Revenue Management (computes daily yields and optimises availability), Disruption Management and Passenger Recovery (compare and apply recovery strategies with optimisation templates) and the SkyWORKS schedule editor, which evaluates business scenarios. | Revenue managementDisruption recoveryNetwork planning | OR · workflow softwarewhyProduct pages name availability optimisation and a recommendation engine; no solver or model named. sources amadeus.com phocuswire.com | Established | reads ops, crew or timetable systems | - | amadeus.com amadeus.com amadeus.com | |
| Flix (network planning, pricing and yield platform) | operator in-house unit | rail | AR6 | Flix's own technology platform plans the FlixBus and FlixTrain network and prices seats: over 495 million seats managed daily by its AI-based system, about 1.2 million rides forecast daily, more than 44 million seats price re-optimised on average every day. Bus-first; FlixTrain is its open-access rail operator. | Revenue managementOffer optimisation and dynamic pricingNetwork planning | ML · ORwhyMachine learning for network planning and pricing (Mediarail); price re-optimisation (flixbus.com). | detailbuyer Flix itself (own operation) | Established | reads ops, crew or timetable systems | ? | corporate.flix.com flixbus.com ledicoferroviaire.mediarail.be |
| FLYR | incumbent vendor | airline | AR6 | Revenue operating system for airlines (forecasting, pricing, revenue management, offer and order retailing) built on deep learning; Air New Zealand adopted FLYR Revenue Management in 2020. | Revenue managementOffer optimisation and dynamic pricing | MLwhy"leveraging artificial intelligence and deep learning technology" for forecasting and pricing (flyr.com press release). sources flyr.com | Established | reads ops, crew or timetable systems | - | flyr.com flyr.com cbinsights.com | |
| PROS (airline offer and revenue management) | incumbent vendor | airline | AR6 | Offer management and revenue management software for airlines; the platform "enables commercial teams to align capacity with demand and coordinate pricing, merchandising and offer strategies", with real-time dynamic pricing and dynamic ancillary pricing. Since the Thoma Bravo acquisition the company focuses on airlines. | Revenue managementOffer optimisation and dynamic pricing | - | detailTake-private valued at USD 1.4bn (Investing.com); not a deal size with an airline. | Established | reads ops, crew or timetable systems | ? | pros.com en.wikipedia.org investing.com |
| Sabre Mosaic (airline revenue and offer management) | incumbent vendor | airline | AR6 | Airline revenue optimisation, offer optimisation and schedule planning products in the Sabre Mosaic suite. | Revenue managementOffer optimisation and dynamic pricingNetwork planning | OR | - | Established | reads ops, crew or timetable systems | 4,650 parent Sabre Corporation: 4,650 employees (2025), Wikipedia infobox https://en.wikipedia.org/wiki/Sabre_Corporation; division headcount not public | sabre.com en.wikipedia.org |
| ÖBB-Personenverkehr (Operations Research team in yield management) | operator in-house unit | rail | AR6 | Three OR specialists support ÖBB-Personenverkehr's yield management group with reporting, programming and forecasting; 2023 main project was an automatic forecast of utilisation rates for day trains. | Revenue management | analytics | detailbuyer ÖBB-Personenverkehr yield management (own operation) | Established | reads ops, crew or timetable systems | 3 workshop invitation: 'a team of 3 Operations Research specialists' | gor-ev.de |
| Volantio | startup | airline | AR6 | Post-booking revenue management: machine learning that moves passengers who volunteer from flights close to selling out to flights with space, freeing seats to sell; Alaska Airlines reports over USD 20m a year in benefit. | Revenue management | MLwhySeries A release: the GreenLeaf platform uses machine learning for demand reallocation. sources volantio.com | Early | reads ops, crew or timetable systems | - | volantio.com volantio.com | |
| Fetcherr | startup | airline | AR6 | Generative AI software for airline pricing and inventory, sold as a 'Market Model' that sets prices and allocates supply in real time. | Offer optimisation and dynamic pricingRevenue management | MLwhyOwn pages: a proprietary neural network 'Market Model' that predicts market moves and sets prices; careers list deep-learning and forecasting data scientists. hiring 14 open tech roles: data science 6, software 4, other 4 fetcherr.io (Page says 38 open jobs; only the first 17 rows were in the page source (Load More not followed). Counted from those 17: data science 6 (incl. director and team lead), software 4 (system architect, 2 DevOps, engineering team lead), other 4 (algorithm team lead, IT team lead, 2 support engineers). Left out: customer success director, payroll manager, ferries revenue-management SME (product). Departments filter shows 'Product & AI (6)'.) | detailbuyer airline pricing and revenue management (testimonials from Director of Revenue Management at Azul, VP Pricing & Revenue Management at Virgin Atlantic, CRO at VivaAerobus) | reads ops, crew or timetable systems | ? | fetcherr.io fetcherr.io | |
| Kambr | acquired startup | airline | AR6 | Airline revenue management startup from Minnesota, a Navitaire partner since 2021, acquired by Amadeus to extend its revenue management offer for airlines. | Revenue management | analyticswhysources amadeus.com | reads ops, crew or timetable systems | ? | amadeus.com | ||
| Optifly | startup | airline | AR6AR4 | Airline schedule optimisation platform (KHIMBA Ai Ltd) for long-term network optimisation, in-season schedule changes, aircraft allocation, slot management and crew reduction, with an agentic AI layer (Opty) that runs and compares schedule scenarios. | Network planningFleet assignment | OR · agentswhyHomepage: proprietary mathematical optimisation at network scale, with AI agents for scenario exploration. sources optifly.com | detailHomepage says products are available individually or as a suite. | reads ops, crew or timetable systems | ? | optifly.com optifly.com |
Group: startup (under about 200 people and founded 2010 or later), incumbent vendor (a vendor over 200 people, listed, or a product line of one; also an established specialist, under 200 people but founded before 2010, marked in the row), operator in-house unit, acquired startup. Problems: the named problem classes the company's own pages say it solves, primary first; hover for the definition. Technical core: the one to three competencies the product stands on, ranked, from open roles, engineering posts and product pages; open "why" for the reasoning, the role counts and the sources. Money signals: deal-size band and what it is per, then delivery model and pricing unit; open "detail" for revenue, funding, buyer, named competitors and sources. Operator ties: large operators or makers the company is tied to and how. A tie is one of six kinds with a large operator or maker in this sector (airlines, rail operators, infrastructure managers, aircraft and rolling-stock makers, airports and air navigation providers; more than 10,000 employees or more than USD 5bn revenue, or a national operator or infrastructure manager of any size) or its venture arm or subsidiary: it invested, a joint venture, the company is a spin-off, a founder worked there, it acquired the company, or a formal partnership with a joint announcement or a co-built product. Customers alone do not count. Left out: accelerator cohorts, multi-company funds, and partner claims that rest only on a logo, a marketplace or ecosystem listing, or the startup's own partner page. Subsector: which part of the sector the company says it sells into, on its own pages: airline, rail, air traffic and airports, or both airlines and rail. A dash means nothing was found in a public source. Amounts in other currencies are converted to US dollars only to place them in a band, at approximately 1 EUR = 1.15 USD, 1 GBP = 1.33 USD, 1 CHF = 1.25 USD, 1 AUD = 0.65 USD, 1 CAD = 0.72 USD, 1 CNY = 0.14 USD, 1 TWD = 0.032 USD, 1 JPY = 0.0066 USD, 1 KRW = 0.00071 USD, 1 INR = 0.0113 USD, 1 CZK = 0.047 USD, 1 NOK = 0.098 USD, 1 SEK = 0.105 USD; bands are rough.
How capacity crosses company walls in airlines and rail, and who owns it
Between organisations is a cross-cutting flag on this map, not a layer: a company whose pages show it modelling parties outside one operator keeps the flag and stays on the layer of the service it sells. 4 of 67 companies carry it: DB InfraGO (ADA-PMB dispatching assistant) (AR3), RMCon International (RailSys) (AR5), SMA und Partner (Viriato) (AR5), R-NEST (EUROCONTROL Research Network Strategic Tool) (AR5).
1. The signal: capacity, declared by one party and allocated by another
In this sector the signal that crosses walls is capacity: a slot at a congested airport, an ATFM regulation on an airspace sector, a train path on a stretch of track. Operators do not set it. They request it, and a third party grants it.
Airport slots. IATA's slot page says slots are used only at Level 3 airports, where demand exceeds runway, terminal, gate, stand or airspace capacity. It counts 160+ Level 2 and 200+ Level 3 airports, and says 50% of all passengers depart from a Level 3 airport [1]. Allocation starts "with the capacity declared by the airport or competent authority", and the coordinator then weighs historic precedence, new-entrant requests and operational feasibility. Airlines keep historic precedence only if they meet the 80% utilisation rule, and new entrants get priority for at least 50% of the slots in the pool [1]. The 80% and 50% figures are as IATA states them on that page.
ATFM and the rolling plan. The EUROCONTROL Network Manager publishes a Rolling Seasonal Plan every Friday, covering eight weeks and consolidating data from 350 airlines, 68 area control centres, 55 airports and 43 States. Edition 277 appeared on 6 November 2026. Operational stakeholders review it every Monday, and the stated goal is that "capacity is available at ACCs ... and at airports, to meet the expected traffic demand" [3]. The Network Manager's main page gives the same figures and says six or eight weeks [4]. The plan began as the NOP Recovery Plan, first published 30 April 2020, when traffic fell to 2,099 flights on 12 April 2020 [3].
Rail paths. The RNE and FTE guideline on conflicting capacity requests (TTR, the redesign of timetabling) describes allocation from capacity model (X-36 to X-18 months) through path planning (X-16 to X-12) to the annual timetable. It cites Directive 2012/34 articles 45 to 47: the infrastructure manager must meet requests as far as possible, coordinate conflicting ones, and declare a section congested if it cannot [5]. The case I found with a scenario step is a new or changed traffic demand, where "an IM creates scenarios for the various options with a focus on the capacity consumption" and the best solution is taken "together" with the applicant [5]. Caveat the earlier candidate note missed: the document's own disclaimer (Feb. 2023) says it "has not yet been released for use in daily business" and "serves only to illustrate potential situations" [5].
2. Who carries it today
- Airports: an independent slot coordinator, not the airport, government or airline. IATA says so [1]; the EU regulation requires the Member State to appoint one who acts "in an independent manner", "neutral, non-discriminatory and transparent", and responsible for slot allocation [2]. Airlines, airports and ATC sit on a coordination committee that advises but does not allocate [2]. ACL describes itself as the world's largest independent coordinator, working with 80 airports, offering coordination, data solutions, special-event coordination, online coordination and consultancy [6]. The rules are the Worldwide Airport Slot Guidelines, edition 4 effective 1 August 2025, developed by airlines, airports and coordinators through the Worldwide Airport Slot Board [1].
- Airspace: the EUROCONTROL Network Manager, a public body with 20+ years of centralised flow management [4]. Its plan is a data consolidation and a weekly meeting with stakeholders [3].
- Rail: the infrastructure manager, under the EU directive, with RNE as the manual the managers use among themselves [5].
3. Who sells there
Near-misses placed elsewhere:
- The page calls it a research tool, "a stand-alone desktop application", for evaluating advanced ATM concepts. It is not sold. The same page says R-NEST shares its base with NEST, "the NM operational tool for capacity planning" [7].
Sources
[1] https://www.iata.org/en/policy/slots : slot definition, Level 2 and 3 counts, independent coordinator, 80% rule, 50% new entrants, WASG edition 4. [2] https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:31993R0095 : Regulation 95/93 (1993 text), Article 4 on the independent coordinator, coordination committee. Not the current version. [3] https://www.eurocontrol.int/publication/european-network-operations-plan-2026-rolling-seasonal-plan : edition 277, 6 Nov 2026, eight-week plan, 350 airlines, 68 ACCs, 55 airports, 43 States, history from April 2020. [4] https://www.eurocontrol.int/network-manager : Network Manager role, six or eight-week rolling plan. [5] https://rne.eu/wp-content/uploads/2023-02-28_A3b-Draft-Guidelines-Allocation-principles.pdf : TTR allocation guidelines, X-36 to X-12 phases, Directive 2012/34 articles 45 to 47, scenario step, draft disclaimer. [6] https://www.acl-uk.org/ : ACL, 80 airports, services list. [7] https://www.eurocontrol.int/solution/rnest : R-NEST research tool, CASA and STAM emulation, shared base with NEST.
method
Companies were found by web search, and each row rests on the sources linked in its last column. The six layers come from the companies' own services. A dash means nothing was found in a public source. A region with few companies means few searches went there, not that it is empty.