Azul scales AI pricing across more than 70% of its network
Azul was the first airline to pilot Fetcherr's AI pricing in 2022, and by December 2025 the supplier said it priced more than 70% of the Brazilian carrier's network. Azul's revenue management director has put the airline's weekly gains from AI at about $6m. SITA's machine learning climb optimisation entered a trial in 2023.
Azul, which SITA describes as Brazil's largest airline by number of flights and destinations, became the first airline to pilot Fetcherr's demand prediction and pricing in September 2022. At the time it lacked a modern pricing platform, according to Fetcherr, which rolled the system out market by market.
In December 2025 André Américo, who runs network and strategic planning and revenue management at Azul, told Aviation Week that AI added on top of the airline's existing systems was producing about $6m a week in gains across revenue management, network planning and maintenance. The published account does not break the figure down by area or explain how it was calculated.
Selected implementations
The work behind the evidence
01 / Deployed
More than 70% of the network priced by AI
The operational problem
Revenue analysts can review only so many markets a day. Fares in the rest follow rules set in advance, which respond slowly when demand shifts.
What the evidence reports
Fetcherr said on 8 December 2025 that its system now prices more than 70% of Azul's network and produces more than 3 million fare recommendations a year. It credits what it calls a Large Market Model, which simulates market outcomes and competitor moves, and says the system has reclaimed more than 2,000 analyst hours with no filing errors.
Operational significance
At that coverage the model sets most of Azul's fares, and the analysts' work moves to exceptions and strategy, as Fetcherr describes it. Pricing is also one of the few airline AI uses where revenue can be compared directly between markets priced by the model and those priced by hand.
Evidence still needed
The coverage, recommendation and hours figures are Fetcherr's. Neither company has published a revenue effect for pricing alone, and Azul's $6m a week estimate covers several areas without a breakdown.
A source can verify an announcement without verifying its projected outcome.
02 / Deployed · limited scope
Climb optimisation on trial
The operational problem
The climb burns fuel faster than any other phase of a flight, and the best speed and acceleration height depend on the individual aircraft and the weather on the day.
What the evidence reports
SITA said in October 2023 that Azul had become the first airline in the Americas to use its Mission Watch weather service for dispatchers, and had agreed a trial of eWAS weather for pilots and OptiClimb across its fleet of 177 aircraft. SITA describes OptiFlight, the suite that includes climb optimisation, as a machine learning tool that recommends climb speeds and acceleration levels from performance models for each aircraft, 4D weather and the flight plan. Daniel Tkacz, Azul's operational vice president, noted that SITA had been its datalink provider since 2009.
Operational significance
Of the three SITA products, only climb optimisation is presented as machine learning. Mission Watch and eWAS are weather services, useful to dispatchers and pilots but not AI as described.
Evidence still needed
The 2023 release describes a trial, and Aerostack has found no announcement that it became a fleet deployment. The 5% average climb fuel saving SITA quotes is a general product figure, not an Azul result.
A source can verify an announcement without verifying its projected outcome.
The wider register
Other tracked activity
Other records are shown separately, including non-AI technology and programmes at different stages of deployment. A register entry does not establish a measured AI benefit.
No other publicly listed records.
Editorial assessment
Azul is one of the clearest airline cases of AI pricing at scale, with a coverage figure from the supplier and an estimate of gains across the whole business from the airline itself.
Separating the two would complete the case: how much of the $6m a week comes from pricing, against which baseline, and whether the climb optimisation trial went fleetwide.
Live data from the Aerostack register. Status reflects published evidence only. See Our approach for how records are graded.
Aerostack register · 2 tracked deployments
Aerostack tracks AI deployments across airports and airlines, with the published evidence behind each record. Photography is real aviation imagery. Record thumbnails may show the named airport or airline; other photographs do not.
