In December 2022, Southwest Airlines faced a massive operational meltdown when over 16,700 flights were canceled following Winter Storm Elliott. Blame was primarily assigned to the failings of Southwest’s crew scheduling software, which was overwhelmed by the number of flight changes necessary. The aging program, named ‘SkySolver’, ended up having to be shut down, and crew schedulers had to manually make flight changes.

The disruptions lasted from December 21 to December 30, and it was the worst operational meltdown in US airline history. Southwest Airlines was fined $140 million by the DOT, the agency’s largest fine in history, while the carrier recorded over $1 billion in losses due to the events. The airline was significantly more impacted by the winter storm than others, and the SkySolver software that created the chaos is only now slated to be replaced by a new cloud-based architecture.

The Failure Of SkySolver

Credit: Shutterstock

As Winter Storm Elliott approached Denver International Airport, Southwest’s largest operating base, the airline began proactively canceling flights at Denver and other nearby airports. As the storm worsened, more flights were canceled at other stations, and Southwest called on SkySolver to manage the disruptions.

First implemented in 2004, the software is intended to create new routings for planes and crews, allowing the airline to quickly revise schedules during operational disruptions. SkySolver is designed to run up to 300 scheduling changes per batch, but Southwest’s daily flights have increased by an average of 69% ever since SkySolver was first implemented.

The system ended up being completely overwhelmed by the number of cancellations, and what’s more, the system could not process changes in real time. Flights were being canceled faster than the software could respond, and SkySolver began sending crews on nonsensical routings while planes had no crews assigned to them. As such, flight cancellations increased, and crews began to time out while being placed on hold with crew scheduling for hours.

With this in mind, the airline’s scheduling team ended up shutting down the system, reverting to manual control. The crisis was an expensive embarrassment for Southwest, which lost customers and took days to recover. There was plenty of blame to go around, with some pointing fingers at the carrier’s point-to-point operations while others looked at the company’s culture, but as the sky cleared, the shortcomings of SkySolver became increasingly apparent.

Ancient IT At Southwest Airlines

Credit: Shutterstock

Concerns had existed for years about the limitations of SkySolver, which were flagged in a 2018 internal report. However, company leadership reportedly ignored warnings to keep costs low, as the carrier continued to record profits and has long been focused on a low-cost operating model. The airline also lacked a backup system, and employees were not trained to respond when everything fails, leaving people clueless about how to remedy the issues. It was a system destined for failure.

Southwest had made numerous modifications to its version of SkySolver since 2004, but the system itself was still designed for an era when routings were simpler, and flight volumes were far lower. Company leadership overlooked the system’s shortcomings because it was always good enough, right up until it wasn’t. Meanwhile, the FAA essentially allowed Southwest Airlines to manage its operational disruption plan on its own, a practice that was quickly ended after the meltdown.

Afterwards, the FAA mandated that Southwest modernize its IT systems and began performing quarterly audits to ensure compliance. The airline was frequently labeled in the media as a ‘cautionary tale’ for other carriers to observe, and Southwest’s reputation with the flying public dived. While Southwest was still profitable in 2022 as well as in 2023, its income was far lower than pre-pandemic, and weak earnings in 2023 made it a target for activist investor group Elliott Investment Management.

The New Software Coming To Southwest

Credit: Shutterstock

On June 17, 2026, Southwest Airlines announced that it was partnering with Amazon Web Services to modernize its IT systems. The carrier currently has a largely on-premises environment, but will transition to cloud-based, AI-agent-enabled architecture. The airline hopes to complete the transition by 2028 and is aiming to make its systems quicker, more flexible, and more reliable. The partnership with AWS will cover a wide range of functions, ranging from customer-facing systems to operations.

AI agents will be deployed across the airline’s functions and teams, while AWS’s Kiro will be used to modernize Southwest.com. Kiro is AWS’s agentic coding service, and it will be leveraged to refactor legacy code much more rapidly. AWS’s AI agents will also enable the airline to build software much more quickly, with engineering teams holding responsibility for the operation of the AI agents rather than doing all the work on their own. They will still hold responsibility for the outcomes.

The move away from on-premises technology towards cloud-based architecture will dramatically shorten development timelines. In addition, Southwest aims to simplify its technology environment and improve cooperation between different systems. Southwest is also utilizing Amazon SageMaker to develop machine learning models, while AWS Lambda will further optimize the carrier’s ability to store pricing data. As a whole, the new partnership will make Southwest a more efficient and flexible airline in all aspects.

Changes To The SkySolver Software

Credit: Shutterstock

SkySolver is still around at Southwest Airlines, but it’s now been rebranded as ‘Crew Optimization’. The airline partnered with General Electric to significantly improve Crew Optimization’s functionality and capabilities. The system’s prior flaws were thoroughly addressed, while Southwest’s crew scheduling and customer phone systems were upgraded to accommodate higher call volumes during surges.

In total, Southwest spent over $1.3 billion on IT improvements in 2023. What makes revising and improving systems like Southwest’s SkySolver/Crew Optimization difficult is the years of continuous changes, sometimes undocumented, that make the system harder to work with.

Often, the engineers assigned to make the changes may no longer be with the company, leaving new employees in the dark on how to further optimize the software. This isn’t an issue unique to Southwest, as this phenomenon occurs with airlines around the world. This makes improvements and optimizations time-consuming, tricky to perform, and expensive.

Leading up to the 2022 meltdown, Southwest’s leadership team was aware that SkySolver was antiquated and unsuitable to handle major operational disruptions, yet the cost of updating or replacing it was eye-watering. As such, leadership left the system largely as is, hoping that it would be enough to handle any future disruptions, and then the airline later spent over $1 billion improving it anyway after an operational disaster.

Technology Problems At Airlines

Credit: Shutterstock

Southwest is hardly the only airline to have faced IT issues. In July 2024, Delta Air Lines flights were severely disrupted following a faulty update by cybersecurity company CrowdStrike. It has since been described as the largest IT outage in history and affected several other industries in addition to the airlines. However, while other carriers quickly recovered, Delta canceled roughly 7,000 flights and experienced system-wide issues for over five days before returning to normal operations.

While Delta pointed the finger squarely at CrowdStrike and Microsoft, it was countered by claims from the latter two parties that Delta’s slow recovery was due to antiquated IT infrastructure. Delta itself has also stated that its crew tracking software had failed, leaving the airline unable to properly assign crews to flights while pilots and flight attendants kept timing out.

The incident was also especially damaging to Delta as its premium brand image is built heavily on reliability, and the incident cost the carrier roughly $550 million. Across the industry, airlines have let their software systems age with minimal updates and are essentially operating on a best-case-scenario philosophy, rather than being prepared for the worst possible outcome.

IT software is expensive to update, and given that operational disruptions aren’t especially common, airlines have generally accepted the risk of letting their antiquated systems remain stagnant while flight and passenger volumes increase. Implementing IT improvements ultimately doesn’t provide much immediate benefit, while the cost can be astronomical.

As such, many airlines look to keep their systems functional during normal operations and record higher profits. Most leaders would rather be remembered for delivering record profits rather than as the individual who spent hundreds of millions on upgrading software systems that were already functional. As such, the industry continues to be crippled by aging IT that only works decently when circumstances are perfect.