The flight plan for AI: How we’re building a culture of innovation at Turkish Technology

When you think of Turkish Airlines, you might picture the vast network of flights that connects more countries than any other airline in the world. You might think of the more than 80 million passengers we serve annually, or the incredible logistical complexity of operating such a massive enterprise across 130 countries and 353 destinations.

At Turkish Technology, it's our job to build the digital engine that powers it all. Our mission is to not only manage Turkish Airline systems today, but to build the next-generation digital products that will define the future of aviation. And for us, that future is undeniably driven by AI.

My team and I didn't approach AI as a collection of one-off pilot projects. Instead, we saw it as a strategic lever to accelerate our entire digital transformation. Our goal was to build a hardened, scalable, and replicable AI infrastructure that could be woven throughout the fabric of the organization. We embarked on this process nearly 3 years ago, and while we’ve made incredible progress, we know there's no end to this journey–the airplanes always need to keep flying to their next stop. It's an ongoing process and a continuous race to adapt as technology–especially generative AI (gen AI)–evolves at a rapid pace. 

Beyond the code: A cultural shift

One of our most important discoveries was that a technological transformation doesn't begin with a software platform. It begins with your people. We realized that our biggest challenge wasn’t technical–it was organizational. Introducing new technologies is one thing, but getting people to adopt them and change their daily workflows is another.

AI projects impact everything: how decisions are made, how responsibilities are defined, and how business units operate. We had to focus on building awareness at the leadership level, helping managers understand the strategic value of AI, and fostering a sense of ownership across the company.

To kickstart this cultural shift, we trained over 200 employees from various business units in data science and AI. These weren't just engineers or data scientists. They were people from finance, operations, and customer service. We then integrated them into real projects, giving them a chance to apply their new skills to the problems they knew best. This was a critical realization. Some of our most significant progress came from business users who understood their domain better than anyone. By empowering them with the right tools and support, we created a more inclusive and scalable approach to AI adoption. This transformed a technical initiative into a cross-organizational grassroots movement.

Delivering real value at scale

Our efforts were not just about enhancing our culture, of course, we needed to deliver real business value. The scale of Turkish Airlines provides both an immense challenge and an incredible opportunity. To make the most of this, we prioritized domains that were both data-rich and operationally impactful.

One of our first major successes was our Tail Assignment Optimization solution. This system uses AI to match specific aircraft to routes based on efficiency, availability, and sustainability goals. The result? Incremental improvements in fuel efficiency that, because of our scale, translate into significant cost savings and move us closer to our sustainability targets.

Our AI journey quickly expanded into other areas:

• Dynamic pricing and cargo revenue management: By using AI to better understand market demand and passenger behavior, we've optimized our pricing strategies.
• Turnaround AI: We use computer vision and airport cameras to monitor ground operations in real-time, helping us reduce delays and improve on-time performance.
• Personalized campaigns: AI helps us move beyond static-fare classes to deliver real-time, tailored offers to our passengers across all digital touchpoints.

Navigating the regulations, securing our future

The aviation industry is one of the most heavily regulated in the world, and in Türkiye, public cloud usage is tightly restricted due to data privacy and compliance requirements. This meant we had to build a strong on-premise foundation for our AI projects. We needed a platform that was flexible and scalable, and that also provided a strong security posture and helped streamline compliance. We chose to partner with Red Hat because their hybrid architecture provided the best solution. It allows us to run our workloads primarily on-premises while giving us the flexibility to scale into the public cloud if needed, without having to redesign our applications.

This focus on a strong, flexible, on-premise foundation is now guiding our next big leap: gen AI. We are building for this future on a privacy-first, enterprise-grade AI infrastructure. We’re not just exploring gen AI—we’re embedding it into our core operations.

We’re developing TK GPT, an internally-hosted large language model (LLM) to empower our employees with tools for documentation, summarization, and software development. We're also building:

• A self-hosted code assistant to accelerate our engineering workflows.
• A contract assistant to help our legal and operations teams generate more secure and consistent documents.
• A gen AI-based chatbot for our call centers to help agents access information more efficiently.

On the passenger side, we’re working on LLM-powered assistants that will help travelers to complete tasks like booking queries and checking flight status updates through conversational commands. All of these systems will run on our in-house GPU infrastructure, allowing us to meet regulatory requirements while continuing to innovate.

Looking back, we’ve learned an important lesson. While a technical transformation starts with platforms, it becomes sustainable only when it reaches every level of the organization. Had we realized this earlier, we would have brought first- and mid-level managers into the data literacy conversation much sooner. We’re doing that now through new training programs, and the internal buy-in and cultural alignment are accelerating.

At Turkish Technology, AI is no longer just an IT initiative. It’s embedded in how we plan and operate, and in how we engage with Turkish Airlines passengers. Now, with a resilient, innovation-driven culture at our core, we are well-positioned to scale this new way of thinking across our entire organization and to help shape the future of aviation.