Fly First, Regulate Later: Why Europe's Rules Are Pushing Drones Abroad

From the Travel Tech Podcast newsletter for episode 25, Fly First, Regulate Later: Why Europe's Rules Are Pushing Drones Abroad.

Julie Garland started as an aircraft maintenance engineer on Aer Lingus heavy checks, went on to fly as a commercial captain, then qualified as a barrister specializing in aviation law before founding a drone training company. That's one guest's résumé, not three. I wanted that perspective in front of those of you who read this newsletter, and I got it, alongside a second guest who's spent her career moving between big tech and the sharpest edges of aerospace.

Sarah Owen started in an entry level procurement job buying spares for an aviation group, wrote for a music magazine on the side to interview her heroes, then spent years selling landing gear at an engineering level before senior stints at Google, Salesforce and Assaia, and now Sutherland while taking on leading roles at the Aviation Club UK Forum and the Royal Aeronautical Society.

Listen on Apple Podcasts, Spotify, and YouTube.

Both guests kept circling back to the same fault line: Europe writes aviation and drone regulation before the technology has been operated at any scale, while the US and China let operators fly first and regulate from what they learn.

The Cost of Regulating Before You Fly

That tension has a structural cause. Europe's drone regulations have been amended and layered for six years, and the current Joint Simplification Board project exists specifically to strip back rules nobody fully re-examined as they piled up. One concrete result: a major European drone delivery operator is now concentrating its expansion in the US, UK, and UAE, not because the technology doesn't work, but because planning permission, noise rules, and airspace access for drone operations still don't exist as a coherent framework at home.

AI is adding a different kind of pressure. Tools like OptiFlight already rebalance an aircraft's climb profile between horizontal and vertical thrust using AI, saving roughly 5-6% on fuel for that phase of flight, a meaningful number at current fuel prices. Where it's working is turnaround optimization and computer vision on the apron. Where it isn't yet is data sharing: airports, ground handlers, and airlines each sit on their own slice of operational data, and commercial incentives often actively discourage pooling it, on top of real, unresolved concerns about AI hallucination in a safety-critical environment.

Julie Garland, CEO of Avtrain, and Sarah Owen, VP for Europe Travel and Transportation at Sutherland, cover all of this and more: the regulate-then-operate divide, gender representation across pilots and maintenance, apprenticeship stigma, and what a genuine safety reporting culture looks like when you're building one from scratch in a young industry.

Key Takeaways

• Regulate-then-operate has a real cost: Europe's habit of legislating before operating at scale is a named reason drone operators are choosing to expand in the US, UK, and UAE instead.
• AI's easiest wins are already banked: Flight-path optimization tools are already delivering measurable fuel savings, in the range of 5-6% on climb.
• The bigger blocker is organizational, not technical: Siloed data across airports, ground handlers, and airlines, often reinforced by commercial incentives, is the harder problem to solve than the AI itself.
• Gender representation has barely moved: Female representation sits around 5% across pilots, ATC, and maintenance, with maintenance closer to 2-2.5%, despite falling technical barriers to entry.
• Safety culture has to be built, not assumed: The drone industry is only now moving toward the kind of non-punitive, mandated incident reporting that took commercial aviation decades to establish.

🔧 Two to Build

This week's guests work at the sharp edge of putting autonomous systems into real operational environments, so the builds follow suit.

Build a computer vision pipeline that tells ops teams exactly where turnaround time is being lost at the gate.

Aircraft sit on the ground longer than they should, and most airlines cannot say precisely why. Camera feeds from the apron, combined with object detection and activity recognition models, can clock every ground handling task in real time and flag deviations before a delay compounds. Sarah Owen built parts of this at Assaia, where the insight was simple: you cannot optimise what you cannot measure, and humans watching monitors do not produce structured data. The models do.

Catch: the hard part is not the vision model, it is getting ground handlers to trust a system that effectively scores their work.

→ Computer vision for aircraft turnaround

Build an automated risk scoring engine that classifies UAS integration scenarios by airspace complexity, not just vehicle weight class.

Current regulatory frameworks sort drones into broad categories that were designed for known aircraft types. Higher risk environments, dense urban corridors, beyond visual line of sight operations over crowds, do not fit those templates cleanly. An automated risk assessment tool that ingests airspace structure, operational parameters, and failure mode libraries can produce a quantified risk profile that regulators and operators can both act on. Julie Garland described exactly this gap emerging from EASA workshops: the certification logic exists, but the tooling to apply it at scale does not yet.

Catch: building for one regulator's risk model means rebuilding when another jurisdiction uses a different baseline, so the data model matters more than the algorithm.

→ Automated UAS risk assessment

Alex Brooker, Host of the Travel Tech Podcast, Founder of Airside Labs