Southwest Airlines is one of the largest carriers in the world. Other than having by far my absolute favorite airplane livery, it's a massive enterprise based in Dallas, Texas, with more than 72,000 employees, over 4,000 daily flights during peak travel periods, roughly 134 million customers annually, and service across 120+ airports in 12 countries.
At this scale, what really matters is operational reliability and speed. Because in aviation, slow operations cause delays, delays cause unhappy customers, and unhappy customers aren't particularly great for the company's revenue. A few minutes of latency in one system can turn into hours of disruption in the real world.
So... can cloud and AI solve it? Absolutely, if done right.
Southwest has selected Amazon Web Services (AWS) as its primary cloud partner to help modernize its technology stack and transform how the airline runs, develops systems, and serves its customers. Through this collaboration, Southwest plans to move away from a predominantly on-premises infrastructure toward a cloud-based, AI and agent-enabled architecture on AWS by 2028.
2028 is not far away from now (Jun 25, 2026). This is very ambitious considering the amount of work that needs to be done.
Moving an enterprise of this size from on-prem infrastructure to the cloud is way more complex than it sounds. It doesn't sound easy either. It's one of the hardest things you can ever do as a software engineer, DevOps engineer, architect, engineering manager, or anyone else involved in it.
It involves a lot of steps, including but definitely not limited to:
And around 10 billion other steps/decisions that you want to make before wrapping things up.
And unlike greenfield systems, airlines don't get the luxury of downtime. Everything has to work while it's being rebuilt. That alone makes airline modernization one of the hardest engineering problems in enterprise software.
AI changes the equation, but not in the simplistic sense of "developers write code faster now." In large enterprises, the real bottleneck is not code generation, it's context. Decades-old systems accumulate implicit knowledge: business rules, edge cases, operational constraints, and undocumented dependencies. Understanding what should change, what must not change, and how everything connects is the hardest part of modernization.
So the real question becomes:
Can AI understand a system well enough to safely change it? That shift, from code generation to system comprehension, is where things start to get interesting.
Also, there are so many important considerations when planning AI agent workflows and their integration into a system, particularly to minimize unintended side effects and maintain efficiency.
Southwest is using Kiro, AWS's agentic development platform, as part of its broader modernization effort, particularly for customer-facing systems and legacy platform refactoring. The goal is not simply to rewrite code, but to systematically evolve large, complex systems into something more maintainable, scalable, and cloud-native, without breaking what already works.
What makes this approach notable is its underlying philosophy.
Kiro is built around SDD: Specification-Driven Development, where you begin by defining what the system should do in a structured, explicit specification before any implementation begins.
That specification then becomes the foundation for:
I'll go deeper into specification-driven development in a separate article, but for now, what I can say is that this is by far my favorite approach to AI workflows. It's significantly more reliable than vibe coding, AI agent looping, or other experimental patterns we see today.
This is especially true for large enterprises with extensive legacy codebases that need refactoring. At a certain scale, systems become so domain-specific and historically layered that even advanced LLMs struggle to reason about them without proper context.
In those cases, the bottleneck is no longer code generation, it's understanding. Without clear specifications, documentation, and well-defined behavior, even the best models start to make unsafe assumptions.
Specification-driven development directly addresses this by forcing clarity upfront. You begin with explicit specs that define behavior, constraints, and expectations before any implementation starts. It forces engineers to understand the system more deeply instead of blindly delegating reasoning to AI. And that's critical, because in complex, mission-critical software, trusting AI without structure is not a strategy, it's a risk. Of course, it's still not 100% deterministic (which it shouldn't be), but it's absolutely much more deterministic than just vibe coding.
In my view, this is one of the most promising approaches for AI-assisted software engineering at enterprise scale. Compared to looser workflows like vibe coding or agent loops, spec-driven development introduces structure and context where it matters most. And that matters especially in environments like airlines, where correctness, safety, and predictability are non-negotiable.
Way to go Southwest, well done!
Southwest's move is not just a cloud migration story. It's a signal of something larger: how enterprise software development itself may evolve in the AI era. If successful, this approach could reshape how airlines design, build, and maintain their systems. Southwest could become one of the pioneers in this shift, modernizing not just its infrastructure, but also decades of legacy software along the way to make its systems more efficient, maintainable, and adaptable.
What makes this partnership interesting is not just what Southwest is doing, but what it might trigger across the industry. Airlines are traditionally cautious adopters of new infrastructure paradigms. These are systems where "if it works, don't touch it" is often the dominant philosophy, because failures are measured in delayed flights, stranded passengers, and operational chaos. So when a major carrier moves this fast toward cloud-native, AI-assisted, spec-driven modernization, it raises a natural question: will others follow?
The direction is clearly positive: better tooling, more automation, and systems that are finally closer to how engineers and businesses actually think.