Travel is full of obstacles. When investors reward scale, the market concentrates. Language still keeps many travelers from local. Even inside our products, A/B testing and AI output can create invisible walls. This question focuses on where the obstacles are still hindering us from retreating and where they start to fall.
Special thanks Digital sardines Newsletters to sponsor this version:
Have you ever felt your Google Ads should work harder? As you prepare for 2026, Digital Sardine, a boutique travel PPC consulting company, will offer free reviews. Gain experts to follow your campaign, discover what works (not), and build a stronger foundation, backed by $500 million in managed ad spend.
Bookings have increased by 158% since January 2021 and are now worth nearly $180 billion in corporate value. MakeMyTrip’s breakthrough has risen, with the Indian market booming, makingmyTrip has risen by 255%. Edreams nearly doubled, Expedia rose 63%, while TripAdvisor fell 39%. Airbnb is outlier: still worth $67B, but its stock has fallen 11%. Investors are beneficial focus and predictable profits. Bookings keep hotels and flights sold at scale, MakemyTrip has been locked in India, both of which have paid off. Airbnb’s lifestyle hub hasn’t convinced Wall Street yet. The trade between TripAdvisor and Edreams is a small part of it, reminding us how centralized global travel has become.
Travelier CEO Noam toister points out what many people already know, but few are willing to say: Some industry reports reflect more of who sponsors them than they actually travel. He mentioned that Skift’s latest 279-page “Travel Status” report covers flights, hotels, car rentals and cruising, but not ground and sea transportation, which is a $198 billion industry that is bigger than cruising, rentals or home rents. Two years ago, Travelier even paid Skift to study the space, which led to a detailed report. Without sponsorship, the industry disappeared. This is surprising given the size and motivation of players like Flix, Blablacar, Omio, Trainline, Travelier and many others who are actively shaping the scale and motivation of players like connecting travel. Read Noam’s full post here.
I asked if some AI models should cover ground and sea transportation. Claude, Openai, Confused, Gemini and Grok all answered some versions of “Yes”. Confused even takes Skift as a source.
If some markets are undervalued, others are undervalued. When we talk about the revolution on the next trip, the focus is usually on the futuristic Agi. Kwon Ping Ho, founder of Banyan Group, believes that the most critical change in travel technology is real-time translation. He compared it to the arrival of low-cost airlines. Once the obstacles drop, the entire market will open. Translation eliminates the friction of language, allowing travelers to connect directly with locals, stay in smaller hotels, and explore without feeling cut off. Airbnb’s Brian Chesky added that AI has not reinvented the platform yet, but in customer service and translation, it is already changing the way people travel. Read +
The recommendation will introduce new discoveries about how travelers use Google’s AI model to plan and book. They will have a deeper look at the results and what players should do for it in the upcoming webinar. Some highlights:
-
56% of travelers went directly to hotels or travel agencies, while Otas hardly reached 6%.
-
Google’s business profile works like a storefront, and when checking reviews, prices and photos before moving on, visit more clicks.
-
The average scheduled time for AI is 104 seconds, nearly three times the 38 seconds booked outside of AI.
-
AI’s response score is high, with an average of 4.3.
Sign up for a webinar here (Hurry, because the registration limit is 500)
Scott Belsky makes a wise point about AI and time. As technology increases productivity every minute, the value of every minute increases. Rather than freeing us, it may prompt us to spend time more intentionally, but something that cannot be copied. This means we will be eager for more human-made travel experiences that feel scarce and personal. He also raised the brand issues in the AI era. If the agent makes decisions for us or more, we don’t care about the brand because when everything looks commoditized, the brand becomes a signal of meaning? Read + lots of time
In AI, “Evals” is to check if the model is doing the test you expect. The equivalent of quality inspection of AI, usability testing or A/B testing. They become essential because the output is probabilistic, not deterministic.
Smart people agree that this is a big deal: Lenny Rachitsky (writer/investor, former Airbnb PM) is called SQL in the AI era. Garry Tan (President of YC) said they are the real moat for AI startups. Sarah Guo (VC, founder of conviction, former general partner of Greylock Partners) calls them new marketing. More information about evals here.
If the AI is going to handle the workflow from support to trip design, we need a way to check if it actually works. In travel, this could mean testing whether AI reliably recommends using safe, bookable routes or hotels that match the traveler filter. The skills to design and run EVAL are as important as building the model itself. This is what separates flashy demonstrations from products that can be expanded.
Mercor.ai founder and CEO Brendan Foody (an AI-powered recruitment platform launched in 2023 and has been estimated at $2 billion).
Alex Bainbridge (founder and CEO of Autoura) published a short paper that a fixed itinerary tour is on the way. He offers 15 ways AI can reshape sightseeing, from real-time adaptation to destinations where you actively manage your visitor process. Human Guides will play more of a coordinated role, stepping in to increase local knowledge or staying on track, while feedback will be achieved, rather than through comments. Itinerary design will be moved from suppliers to travelers who work with AI. He also predicts that the economics of sightseeing will change, with micro-trading replacing fixed tour tickets. Read + Alex Bainbridge
If AI reshapes travel, it is reshaping the business model, too. Lenny Rachitsky pointed to a new model: Saas is hitting growth with full integration into AI in the later stage. Startups may be early adopters of AI, but mature companies show that it can be used as a second move. The next curve usually comes from embedding AI into areas where growth slows.
During the trip, Expedia begins to showcase its appearance. Scout drives $6B incremental partner revenue, AI filter conversions 1.3 times, while flight price forecasts improve conversions by 15%. These are meaningful results on a large scale, but they look more like optimization than resets. I would love to hear what you see.
For start-ups, the perspective is different. This Harvard Business Review The report notes that entrepreneurs aiming to grow rapidly are those who work hard to lean towards AI. For lean teams, it’s a way to go beyond weight…automaticize GUNT’s jobs and test the market faster without adding to the staff they can’t afford. For Lean teams, AI resets less than resets, but reaches scale earlier.
On the other side of the spectrum, it’s worth remembering what to optimize. The A/B test tells you which button can click more, but it can’t tell you what’s worth building. Endless optimizations promote hidden expenses and turn hidden gems into overcrowded docks. At some point, you have to stand up. This is how you build products that people like rather than just convert.
→Explore All 1,358 public roles Now on the Travel Technology Essentialism Working Committee.
📩 For the latest roles, please subscribe Travel Technology Work Newsletter
If you are a startup looking to raise a round (from seed to series D), I can help (free). Travel Investor Network is a private platform where I recommend innovative travel startups to investors and innovators. If you are interested, please first Complete this form.
If you like travel tech essentialist, consider sharing with your friends or colleagues. If you have not subscribed yet, please join us:
And, as always, thank you for trusting my inbox.
Mauricio Prieto
