Beyond the hype: aviation's urgent need for strategic AI implementation.

The data speaks volumes. As stated by Rafa Mercado, Vice President and Consumer & Travel Consult Partner for Kyndryl, a staggering 86% of leaders believe their AI is ‘best-in-class’, but only 42% report seeing a positive ROI. This gap is the industry's most urgent technical debt. 

Everyone I spoke to and listened to at the World Aviation Festival event feels nervous but hopeful about the technology. We must embrace the unknown to grow, even though fear is a normal part of change. Our collective anxieties make us question many aspects, such as:

• How will we change our perspective on data analysis to discover new patterns?
• Will AI truly help us negotiate better contracts and predict maintenance needs?
• Can engaging with next-gen chatbots truly transform customer support, or just complicate it?
• How will we ensure we aren't just selecting a "flashy use case" that ultimately delivers zero business value?

The airline industry has to adapt to the cold, hard fact that the path to ROI requires more than just buying a new system. As Mecardo explains, when proofs of concept (POCs) and pilots are scaled up into production, organisations often confront the challenge of having a lot of technical debt, legacy systems, and fragmented silos of data. Because of these issues, they may not have the necessary computer resources, security posture, or resiliency to take things into production. The airline industry must meet these demands aggressively, tackling technical debt and fragmented data silos - the shaky sandcastle foundation many aviation companies are trying to build upon.

AI is a tool for people. Success hinges on the cultural aspect of companies and industries: enabling and empowering the frontline employees and users who use the system. 

Technology should enhance productivity and decision-making, not immediately replace headcount. While AI is excellent for autonomous, "business as usual" tasks, complex disruption demands human intervention and empathetic communication.

Just a few short years after mastering generative AI, we are already moving towards agentic AI. Think about how a Large Language Model (LLM) combines multiple building blocks of language, similar to how various words create sentences that can be rearranged to change the context. Agentic AI is that, but for control towers. This technology promises to create "autonomous control towers" capable of generating personalised recovery plans for passengers during disruption in milliseconds.

The industry needs to continuously update its strategies and technology to handle the millions of data points needed for hyper-personalisation in dynamic pricing and offer management. They can achieve this only by embracing the speed of change and challenging themselves to get the implementation right from all angles.