Airlines Don't Need Another AI Strategy — They Need Someone With Authority to Run It

An analysis published on the World Aviation Festival's blog on September 23 argues that airlines' AI efforts are stalling not for lack of strategy documents, but because most carriers lack an operating model that assigns clear ownership and accountability once AI systems move from pilot projects into live production. The piece, credited to consultancy DataArt, contends that the shift from advisory AI tools to autonomous agents that act directly on operational systems makes this governance gap increasingly consequential, since mistakes can no longer be caught by a human reviewing a recommendation before it takes effect. Its central recommendation is a centralised coordination function — something like an AI Centre of Excellence — that holds genuine authority rather than an advisory seat: control over which use cases get built and in what order, enforced technical guardrails built into delivery pipelines, named ownership of data quality, and compliance checkpoints triggered automatically by regulatory requirements. The analysis argues this function needs "real teeth," holding the project intake queue, the release gate and an actual budget rather than merely convening meetings. It cites BCG research across more than 1,000 AI implementations finding that returns depend roughly 70% on people and process changes and only 10% on the underlying algorithms, and separate McKinsey findings that fewer than one in ten enterprises have successfully scaled AI agents despite widespread experimentation. As practical illustration, the piece suggests airlines should prioritise unglamorous use cases like fuel invoice reconciliation, where the existing manual workload is large and well understood, ahead of higher-profile but less-proven projects like customer-facing chatbots.