Airline Revenue Management

Part of the worlds-biggest-software-project initiative.

An AI-native, open alternative to legacy airline revenue management systems, built for real-time dynamic pricing and network-level inventory optimisation.

Airline Revenue Management is a candidate platform for continuously balancing seat inventory, fares, and overbooking across thousands of origin-destination markets. It targets carriers underserved by enterprise incumbents — regional, low-cost, and start-up airlines — while offering full-service carriers a modern, classless, AI-driven alternative to retrospective rule-based optimisation.

Why Airline Revenue Management?

Incumbent platforms (Amadeus Network RM, Sabre Mosaic, PROS) are deeply coupled to specific PSS stacks and impose high licensing cost and long 12–24-month feature cadences, leaving mid-size and regional carriers priced out.
Traditional systems still rely on discrete fare-class buckets and historical demand curves, which struggle to respond to rapid demand shifts and competitive fare changes.
A 2026 industry survey found 75% of airline executives identify dynamic pricing as their top revenue growth priority, signalling broad dissatisfaction with fare-class-based approaches.
The only meaningful open-source option, RMOL (LGPL), implements only classical EMSR/LP algorithms with no ML/AI capabilities, no real-time pricing, and no modern retailing support.
AI-native challengers like Fetcherr and FLYR are proprietary SaaS with patent-protected approaches; there is no open AI-native equivalent for the rest of the market.

Key Features

Demand Forecasting & Inventory Control

Time-series and ML-based booking-curve modelling per flight and route
Booking-class availability optimisation using EMSR or LP-based bid prices
Bayesian updating capturing price elasticity, seasonality, competitive signals, and special events
Forecast bias detection and automated re-calibration

Continuous & Dynamic Pricing

Classless, continuously computed seat prices reflecting current demand and willingness-to-pay
Real-time bid-price API for per-shopping-request pricing at low latency
Leg-level and origin-destination level pricing decisions across multi-leg itineraries
Automated open/close decisions for booking classes during transition periods

Overbooking & Group Management

Probabilistic no-show, cancellation, and go-show models per flight and market
Optimised booking limits maximising expected revenue while capping denied-boarding cost
Group block accounting and codeshare inventory handling
Ancillary upsell contribution tracking

Analyst Tools & Reporting

Exception-based analyst workbench with drill-down into underperforming markets
Load factor, yield, spoilage, spill, and RASK dashboards
What-if scenario modelling and back-test environment for RM policy evaluation
Override workflows for revenue analysts

Distribution & Integration

PSS connector architecture for Amadeus, Sabre, and Navitaire
ATPCO fare filing API integration
NDC API support for modern retailing distribution
REST APIs for pricing and distribution systems
Real-time availability updates across GDS, NDC, and direct channels

AI-Native Advantage

AI replaces the rule-based core of legacy RM with continuously adaptive models: deep learning demand forecasting handles irregular booking patterns where ARIMA and exponential smoothing fail, and reinforcement learning can adjust overbooking limits dynamically rather than relying on static historical show-up rates. A real-time willingness-to-pay inference layer enables genuinely classless pricing per passenger per session, while a "Glass Box" explainability layer surfaces human-readable justifications for each pricing decision — addressing the regulatory and analyst-oversight concerns that block fully automated pricing today.

Tech Stack & Deployment

The platform targets a PSS-agnostic, API-first architecture so airlines on Altéa, SabreSonic, Navitaire, Radixx, or Hitit can adopt it without full stack migration. Bid-price calculations must return in milliseconds across global GDS and direct channels, requiring robust low-latency APIs and continuous ingestion of bookings, competitive fare feeds, and event data. Open standards include IATA NDC and ONE Order XML schemas (published, no licence fees) and ATPCO fare filing pipelines. Self-hosted and cloud deployment are both expected; model governance tooling for drift monitoring and human override is treated as a first-class requirement.

Market Context

The airline RM software market is dominated by Amadeus, Sabre, and PROS, which serve major full-service carriers via deep GDS and PSS integration. AI-native challengers such as Fetcherr (2026 BIG Innovation Award; ~$80M USD revenue impact at Azul over three years) and FLYR (10% seating ancillary uplift at Virgin Atlantic) are pressing incumbents to accelerate AI roadmaps. Primary buyers are airline commercial and revenue management departments; the underserved segment is the long tail of regional, low-cost, and start-up carriers priced out of enterprise platforms.

Project Status

This project is in the research and specification phase.
Contributions, feedback, and domain expertise are welcome.

Contributing

We welcome contributions from developers, domain experts, and potential users. See CONTRIBUTING.md for guidelines.

Important: All contributions must be your own original work or clearly attributed open-source material with a compatible licence. Copyright infringement and licence violations will not be tolerated and will result in immediate removal of the offending contribution. If you are unsure whether a piece of code, text, or other material is safe to contribute, open an issue and ask before submitting.

Licence

Licence to be determined. See discussion for context.