Nobody Would Fund a Human for This

From the Travel Tech Podcast newsletter for episode 26, Nobody Would Fund a Human for This.

I've now recorded enough of these conversations that I stopped believing in coincidence. Four founders, none of whom have met each other, working in hotel distribution, airport maintenance, call center infrastructure, and hospitality tech, all independently built the same tool. To those of you reading this across aviation, travel, and AI, I wanted that pattern on the show before I moved on to anything else.

One runs a reconfirmation service that calls hotels directly to verify bookings before guests ever show up, catching the roughly 2% that quietly fall through the cracks of modern distribution. Another built a failure engine inside airport maintenance operations that turns fragmented, paper-based frontline data into something an AI can actually reason about. Listen on Apple Podcasts, also streaming on Spotify and YouTube.

The tension underneath all four conversations: this software was always a good idea, but never one a company could justify staffing with people, until the cost of a call collapsed to a few cents.

The gap between a demo and a departure manifest

That tension has a structural cause. Hotel switches and travel gateways make shopping easier for the traveler by pooling supply across sites, but every additional hop between the OTA, the bed bank, and the property is a place where a confirmed booking can silently disappear, which is how you land at roughly 2 in every 100 hotel reservations going missing by the time a guest checks in.

AI is adding pressure from the cost side, not the capability side. One call center processing up to 500,000 calls a day found that anonymized transcription and analysis now runs 5 to 10 cents per call against roughly $200 in support costs saved, and it's working well where the task is narrow and auditable, less so wherever a customer senses they're talking to a machine and has no way to reach a person.

Mike Putman of Custom Travel Solutions, Anne Marie Pellerin, Yagub Rahimov, and Fred Bean of Hotel Port cover reconfirmation calls, airport equipment failure alerts, gateway-based data anonymization, and whether voice or text will actually win the interface war.

Key Takeaways

• The reconfirmation gap is structural, not incidental: Every additional layer in modern hotel distribution (gateways, hubs, bed banks) adds a point where a booking can get lost before the guest arrives.
• The economics flipped, not the concept: Processing a call for 5 to 10 cents against $200+ in saved support costs is what makes AI phone calling viable now, not a new capability that didn't exist before.
• Data fragmentation blocks predictive maintenance more than AI does: Frontline paper logs, call center software, IT systems, and vendor platforms all sit in silos, and orchestrating them matters more than the AI layered on top.
• Anonymization can be an architecture decision, not a rebuild: A gateway approach can strip private data before it reaches an LLM using a single base URL swap, according to Yagub Rahimov.
• Voice adoption depends on the exit, not the entrance: Customers tolerate AI on the phone only when there's an obvious, fast way to reach a human, and some in the space expect text to overtake voice as the primary channel long-term.

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