Azul Airlines AI Network Planning

$6M Weekly Revenue Gains with AI-Driven Network Planning

Discover how Azul Airlines achieved $6M weekly revenue gains using AI-driven network planning with Double ML, causal inference, and real-time financial forecasting.

THE CHALLENGE

The problem.

Airlines make billion-dollar decisions about which routes to fly, how much capacity to deploy, and what prices to charge. Traditionally, these decisions rely heavily on executive intuition and quarterly financial models that arrive 2-3 months after the fact. By the time you understand what happened last quarter, market conditions have already shifted.

Azul Airlines, Brazil's third-largest carrier, faced exactly this challenge. Network planning decisions lacked quantitative validation. Financial visibility lagged by months. The company needed to answer fundamental questions: which new routes would actually be profitable, how to price tickets when demand patterns shift, and what the real-time financial impact of today's decisions actually was.

The core problem was that Azul could see correlations in historical data but couldn't validate causal relationships. Would adding a flight to a new market generate profit, or would it cannibalize existing routes? Traditional business intelligence tools couldn't answer this. The solution required causal inference, not just correlation analysis.

The initial team working on AI solutions was just 5 people facing resistance from decision-makers who relied on decades of industry experience, creating a trust gap that was as much a cultural challenge as a technical one.

THE SOLUTION

What we built.

Ticket-to-Financial Clarity: Eliminating the Reporting Lag

The first system we built converts ticket sales data into a comprehensive financial picture using live data from Azul's operational systems. Previously, finance teams waited for quarterly models to understand profitability. The new system eliminated that 2-3 month lag entirely, giving leadership real-time visibility into the financial impact of pricing decisions, route changes, and capacity adjustments.

The system includes a PNL forecast correction module that uses real-time data to adjust profit and loss forecasts dynamically. Instead of relying on stale quarterly assumptions, the model updates daily as new ticket sales, fuel prices, and operational costs flow through.

We built the system on AWS and Snowflake infrastructure to handle Azul's operational data volume. The key innovation was speed. By eliminating the months-long delay in financial visibility, Azul could test pricing strategies systematically, increase fares in specific markets, observe the real-time revenue impact, and adjust quickly if demand dropped. This systematic approach contributed directly to the 25% fare increases Azul achieved.

Double ML Causal Inference for Network Decision Validation

Correlation doesn't equal causation, and that's the core challenge in airline network planning. Azul could see that certain routes performed well historically, but couldn't answer the counterfactual question: what would have happened if capacity had been allocated differently? Would a new route generate incremental revenue, or just shift passengers from existing flights?

We implemented Double ML (Double Machine Learning) to extract causal information from correlational data. Double ML combines machine learning with econometric causal inference to estimate treatment effects while controlling for confounding variables like seasonality, competitive pricing, economic conditions, and network effects.

This enabled quantitative validation of network planning decisions. Instead of debating whether a new route would work based on intuition, stakeholders could see model predictions with confidence intervals. The causal inference system became the foundation for the Markets Ranked by Overall Score product, which ranks thousands of potential airline markets for capacity allocation using ML-driven scoring that accounts for causal relationships, not just historical correlations.

In the first two weeks after deployment, the Markets Ranked system generated 1M-2M BRL by identifying optimal network expansion opportunities that weren't visible through traditional analysis.

Ensemble Modeling: 9 Models for Robust Recommendations

No single model is perfect. Different modeling approaches capture different aspects of airline revenue dynamics. We built an ensemble system combining 9 models: 3 prediction models and 3 causal models, each with variations. The prediction models forecast demand and revenue using time series analysis, gradient boosting, and neural networks. The causal models estimate treatment effects using different econometric approaches.

The ensemble combines these models to produce robust revenue predictions and network optimization recommendations. When models agree, confidence is high. When they diverge, the system flags uncertainty and identifies which factors are driving the disagreement. This uncertainty quantification was critical for building trust with leadership, who needed to understand both upside potential and downside risk.

This ensemble approach enabled Azul to identify 500+ profitable flights in non-obvious markets during high season. These weren't routes that traditional analysis would have prioritized. The models identified market opportunities where demand patterns, competitive dynamics, and network effects aligned favorably.

The system runs daily on AWS infrastructure, processing updated operational data and producing fresh recommendations. We built comprehensive monitoring and validation pipelines to catch data quality issues, model drift, and prediction anomalies before they impact business decisions.

Building Trust: From 5 Skeptics to Organizational Buy-In

The hardest part wasn't building the models. It was changing the culture. When we started, 5 people at Azul were actively using the AI tools. Executives had decades of airline industry experience and trusted their intuition.

We took a systematic approach to building trust. We started with validation rather than recommendations, using the causal inference system to explain why certain routes performed well or poorly, confirming or challenging existing hypotheses before proposing anything new. Every recommendation included explainability: why did the model rank this market highly, what factors drove the prediction, and what was the confidence interval.

Delivering quick wins mattered. The 1M-2M BRL generated in the first two weeks from the Markets Ranked system demonstrated immediate value and built credibility. When models disagreed or confidence was low, we said so. This honesty built trust more than overpromising would have.

By the end of the 9-month initial project phase, the user base had grown from 5 to over 10 stakeholders, with network planning decisions increasingly incorporating model recommendations alongside executive judgment.

HOW IT WORKS

The details.

Cutting the Reporting Lag From Months to Instantly

Finance teams at Azul used to wait months to understand whether a pricing decision had worked. We built a system that converts live ticket sales into a full financial picture every day. Leadership can now see the impact of pricing changes, route adjustments, and capacity decisions as they happen rather than in a quarterly report. This shift from waiting to knowing enabled Azul to test pricing strategies systematically and adjust quickly when something was not working.

Answering the Counterfactual Question

Knowing that a route performed well is useful. Knowing whether it would have performed well regardless of your decision is more useful. Traditional analysis cannot answer that question. We used a two-stage statistical technique that separates the effect of a decision from outside factors like seasonality, competitor behaviour, and economic conditions. This let Azul's planning team validate network decisions with confidence intervals rather than gut feeling.

Nine Models Working Together for More Reliable Forecasts

No single model is perfect. We built an ensemble of nine models that each capture different aspects of airline revenue. When the models agree, confidence is high. When they diverge, the system flags uncertainty and identifies what is driving the difference. This approach helped Azul find over 500 profitable flights in markets that traditional analysis would have overlooked.

Building Credibility Before Proposing Anything New

When the project started, five people at Azul were using the AI tools. Executives with decades of experience were sceptical. We spent the first phase validating what the team already believed, using the causal analysis system to explain why certain routes had performed well or poorly. Every recommendation came with an explanation and a confidence range. Quick early wins built credibility. By the end of the initial nine months, over ten stakeholders were incorporating model recommendations into their decisions.

OUTCOMES

What shipped.

  • $6M weekly revenue gains across AI strategy initiatives (reported by Aviation Week, February 2026)
  • 25% fare increases through systematic data-driven pricing optimization
  • 500+ profitable flights added in non-obvious markets
  • 2-3 month financial reporting lag eliminated, replaced with real-time visibility
  • 1M-2M BRL generated in first two weeks from Markets Ranked system
  • Active user base grew from 5 to 10+ stakeholders within 9 months

KEY TAKEAWAYS

What we learned.

  • Start with real-time visibility before optimization. Eliminating Azul's 2-3 month reporting lag enabled systematic testing of pricing strategies that led directly to 25% fare increases.
  • Use causal inference, not just correlation, for high-stakes decisions. Double ML enabled Azul to validate network planning decisions quantitatively, replacing intuition with data-driven counterfactual analysis.
  • Ensemble modeling increases reliability for production systems. Combining 9 models provided robust recommendations and identified 500+ profitable flights that single models would have missed.
  • Build trust gradually through validation, not just recommendations. Starting with 5 skeptical stakeholders, Azul grew to 10+ active users by validating existing decisions before proposing new ones.
  • Quick wins accelerate organizational buy-in. The 1M-2M BRL generated in the first two weeks from the Markets Ranked system demonstrated immediate value and built credibility for broader adoption.
  • Admit uncertainty to build long-term trust. Showing confidence intervals and flagging model disagreement proved more valuable than overpromising, establishing AI tools as reliable decision support.
  • Daily model updates matter for operational decisions. Running models on fresh data daily meant network planners worked with current insights, not stale quarterly reports that lag market conditions.

IN SUMMARY

Bottom line.

In summary, Azul Airlines transformed from flying blind on network decisions to making data-driven choices backed by causal inference and real-time financial visibility. As a result, the $6M weekly revenue gains, 25% fare increases, and 500+ new profitable flights demonstrate the business impact of production AI systems built for operational decision-making, not just analysis.