50 Years of Tech Debt, AI Enters the Chat

From the Travel Tech Podcast newsletter for episode 16, 50 Years of Tech Debt, AI Enters the Chat.

I've covered AI and legacy systems on this podcast before, but I'd never asked someone who actually lives inside the distribution layer to explain how it works, and why it's such a challenge. Knowing this newsletter reaches people across aviation, travel tech, and AI, this felt overdue. This week I did.

My guest has spent his entire career in travel, from corporate travel agent to GDS insider to modern retailing platform builder. He walked me through all of it.

Apple Podcasts, also streaming on Spotify and YouTube.

What emerged wasn't just a history lesson, it was a map of a genuinely unsolved problem: who provides the trust and authentication layer that the AI agent era requires?

What's broken and why it matters now

The GDS solved fragmentation and worked well for decades (although expensive). NDC arrived to modernize retailing and, instead of replacing the old infrastructure, stacked on top of it. Travel agencies now run parallel support for two incompatible standards while according to some, NDC still can't reliably handle refunds or interline itineraries. Airlines are transacting on 3.5 to 4% margins, roughly the profit of a single cup of coffee per ticket, and being asked to re-platform.

AI shopping agents don't help. They shop, compare, and rebook around the clock without stopping. Consumer look-to-book ratios already run 400 to 500 to one. Multiply that by billions of bots that never sleep and the infrastructure math breaks.

Jim Hetzel of TWAI joins me to cover NDC's backwards compatibility issues, the orchestration layer the industry actually needs, the trust gap no one has solved, and how TWAI is enabling modern retailing today without waiting for standards to catch up.

Key Takeaways

• NDC is a messaging standard, not a retailing platform: Becoming a true retailer still requires CRM, personalization, and revenue system investment. NDC delivers none of that.
• Standards incompatibility compounds over time: NDC versions aren't backwards compatible, so early adopters face multi-million dollar re-implementation decisions at every major release.
• AI can interpret what the industry can't agree on: AI sits on top of both legacy GDS and modern NDC simultaneously, normalizing content without waiting for 900 airlines to converge on a standard.
• The trust gap is the unsolved problem: IATA certifies agents and airlines as legitimate counterparties. No equivalent exists for AI agents. Nothing tells an airline whether a bot represents a real buyer or a competitor's scraper.
• AI kills weak intermediaries, not intermediaries: The players who survive will be the ones who solve orchestration, trust, and content normalization at scale.

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