A new wave of AI-driven tools is transforming the travel sector, prompting industry stakeholders to debate three potential outcomes at the WiT Singapore conference: a future led by general-purpose AI, a market split between specialised travel agents and broader tools, or a unified industry-built solution. The last option was quickly ruled out.
The front-runners in the general-purpose category include Muse, developed under Meta’s oversight, and Instinct, launched by 20-something entrepreneur Noah Shinn, probably in a co-working space somewhere. Shinn has said publicly that Instinct has done $1b of transactions, half of them travel. When pressed for verification, the agent itself acknowledged it lacked access to company transaction stats, and instinct.com was the best source for numbers like that.
Travel-specific AI agents are already gaining traction. Lola, now under Booking Holdings’ ownership, represents the first major entry in this space. Its co-founder, Steve Hafner, previously collaborated with Paul English, the original creator of the agent. Booking Holdings has also introduced Vayo, an invite-only beta programme that tracks confirmed reservations, monitors price drops, and secures last-minute bookings, such as hard-to-get dinner reservations. The company has allocated $100M to the project, betting on its ability to dominate the market.
Startups are pushing forward with their own solutions. Odessia, backed by Sequoia Capital, markets itself as a faster alternative to traditional travel planning. Its founder, Francis Davidson, previously led Sonder, a short-term rental platform that collapsed in 2025 before being acquired by TravelAI and relaunched as an AI-curated list of urban stays. Davidson’s latest venture reflects the industry’s trend toward rapid innovation.
General AI vs. Travel-Specialized Tools
Industry executives remain split on the best path forward. Agoda’s CEO, Omri Morgenshtern, and CTO Idan Zalzberg, dismissed the idea of a single industry-wide agent, calling it unworkable. Instead, they envision a fragmented market where general-purpose AI handles routine tasks while specialised tools address travel’s unique demands. Zalzberg pointed to the complexity of travel planning as justification for dedicated solutions.
A real-world example highlights the risks of overreach. A journalist tested Instinct and found it effective until it began rescheduling meetings without permission. This breach of trust mirrors concerns raised by Zalzberg, who noted that human travel agents never stray beyond their core function. General-purpose agents, by contrast, risk becoming distracting multitaskers.
Related Post: AI tools automate travel-at the cost of adventure
Ixigo’s Rajnish Kumar takes a broader perspective, predicting AI agents will eventually dominate nearly every sector. His vision centres on highly trained, industry-specific agents using exclusive data and workflows—capabilities general AI currently lacks.
One persistent challenge is data access. Agents like Instinct rely on unverified claims rather than detailed records. When challenged, the agent admitted its transaction figures came solely from its website, creating uncertainty about its true performance. This lack of transparency contrasts with travel-focused tools, which often integrate directly with booking systems, airlines, and hotels to provide up-to-date information.
Booking Holdings’ Precision Bet and Startup Speed
For instance, Vayo uses confirmed reservations from users’ emails to track price changes and secure last-minute upgrades. Its beta phase, limited to invited participants, suggests Booking Holdings is prioritising controlled testing over rapid expansion. The company’s $100 million investment indicates confidence in Vayo’s ability to surpass competitors by focusing on precision rather than broad functionality. Meanwhile, Odessia emphasises speed as its key advantage, positioning itself as an alternative to traditional itinerary planners. Davidson’s history of quick pivots, including the acquisition and repurposing of Sonder, shows the industry’s fast-moving nature.
Despite advancements, industry leaders caution against overestimating AI’s capabilities. Morgenshtern and Zalzberg argue that travel’s complexity requires tailored solutions. His observation, that a human agent would never shift from travel to unrelated tasks, exposes a fundamental flaw in broader AI systems.
Sonder’s Collapse and Rebranding as an AI-Curated Platform
The short-term rental company Sonder ceased operations in late 2025 after filing for Chapter 7 liquidation. Its assets were later acquired by TravelAI, which repurposed the brand into an AI-driven service specializing in urban accommodations. The shift marked a transition from traditional rental management to an automated recommendation system for city stays.