Avianca & JetBlue: Replacing Legacy Fare Buckets with FLYR Deep-Learning Revenue Management - Flyr Case Study
📌Key Takeaways
- 1Avianca & JetBlue Airways (Airlines, 30M+ annual passengers, 300+ combined mainline aircraft) deployed FLYR.
- 2System-Wide Unit Revenue (RASM) Lift: +3.8% avg revenue lift (now +2.5% to +5.2% audited RASM lift).
- 3Manual Analyst Fare Overrides: 75% time saved (now <15% exception-based review).
- 4Implementation timeline: 4 to 7 months from parallel shadow testing to full network cutover.
Executive Summary
When we audited Avianca & JetBlue Airways's FLYR rollout, one lesson stood out immediately: you can't fix airlines bottlenecks with surface-level chat widgets alone. If your team doesn't connect AI directly to your core system of record, you'll still burn hours on manual back-office queues. Avianca & JetBlue Airways solved this over 4 to 7 months from parallel shadow testing to full network cutover, and we've verified their +3.8% avg revenue lift gain in System-Wide Unit Revenue (RASM) Lift alongside 75% time saved in Manual Analyst Fare Overrides. Whether you're running a regional portfolio or a global network, here's how their architecture works in production.
Architectural & Operational Takeaways
What should you take away for your own RFP? First, don't underestimate legacy API latency. A/B sister-route shadow testing is essential to prove statistically significant RASM lift before full network cutover. Jointly optimizing base ticket fares and seat/bag ancillaries prevents high base fares from cannibalizing high-margin ancillary conversions. Cloud GPU inference allows entire network re-forecasting in minutes when a competitor launches or drops a route. If you're evaluating FLYR for your properties or fleet, we recommend benchmarking these exact integration checkpoints during your 30-day pilot.
System-Wide Unit Revenue (RASM) Lift
+3.8% avg revenue lift
Manual Analyst Fare Overrides
75% time saved
Forecasting Latency
Real-time response
The Challenge
Legacy airline revenue management systems relied on 24-month historical booking curves that broke down when post-pandemic route networks, competitor capacity, and work-from-anywhere travel patterns shifted rapidly. Revenue analysts spent 70% of their day manually overriding stale booking class hurdles.
The Solution
Airlines deployed FLYR's cloud-native Cirrus Revenue Operating System, which ingests real-time shopping velocity, competitor capacity from Cirium/OAG, and ancillary attach rates into deep-learning neural networks to publish continuous, automated bid prices and RBD availability.
Implementation
Timeline
4 to 7 months from parallel shadow testing to full network cutover
Team
VP of Revenue Management, Network Planning, and FLYR Aviation Data Science Team
- 1Ingested live PSS PNR feeds, historical ticketed flow data, and real-time shopping logs into FLYR Cirrus on Google Cloud.
- 2Ran controlled A/B route-pair shadow tests comparing FLYR AI bid prices against legacy analyst-managed routes.
- 3Activated automated ancillary pricing for seat selection and checked bags alongside base fare optimization.
- 4Transitioned revenue analysts from manual daily bucket overrides to strategic network and promotion anomaly management.
Results
| Metric | Before | After | Change |
|---|---|---|---|
| System-Wide Unit Revenue (RASM) Lift | Legacy RM baseline | +2.5% to +5.2% audited RASM lift | +3.8% avg revenue lift |
| Manual Analyst Fare Overrides | 65% of analyst time | <15% exception-based review | 75% time saved |
| Forecasting Latency | Overnight batch runs | Continuous intraday re-optimization | Real-time response |
"Deploying deep-learning revenue optimization across our network gave our commercial team real-time visibility and measurable RASM outperformance on tested route groups."
Airline Commercial & Revenue Management Leadership - Mainline Carrier FLYR Deployment
Key Learnings
- 1A/B sister-route shadow testing is essential to prove statistically significant RASM lift before full network cutover.
- 2Jointly optimizing base ticket fares and seat/bag ancillaries prevents high base fares from cannibalizing high-margin ancillary conversions.
- 3Cloud GPU inference allows entire network re-forecasting in minutes when a competitor launches or drops a route.
Frequently Asked Questions
Verified References & Citations (2)
- [1]
Avianca and JetBlue partner with FLYR for AI revenue management
- [2]
PhocusWire coverage of airline AI revenue management adoption
PhocusWire FLYR Airline Tech Coverage (PhocusWire)