Kayak Co-Founders Launch Lola, AI Travel Agent Backed by Booking Holdings

Kayak Co-Founders Unveil AI Travel Agent with Booking Holdings Support

Lola, a new AI travel agent developed by Kayak co-founders Steve Hafner and Paul English, launched Tuesday with backing from Booking Holdings, the parent company of Booking.com, Priceline, Kayak, and OpenTable1

1

. The AI-powered travel service aims to bridge the gap between conversational AI recommendations and actual bookings, addressing what Hafner identifies as a critical weakness in current large language models. "LLMs do a really good job of answering free-form questions," Hafner explained, "but where they fall down is once they make you a recommendation"1

1

. The platform's name plays on longitude and latitude, positioning itself as a comprehensive solution for the non-flight aspects of travel planning.

Conversational AI Delivers Personalized Travel Recommendations with Real-Time Booking

Lola offers personalized travel recommendations, live availability and pricing, and direct booking capabilities all within the same conversational flow2

2

. The platform can recommend and reserve hotels, restaurants, events, activities, or entire trips through natural language processing. Hafner's pitch centers on combining "all the goodness of an LLM's recommendations with all the goodness of live prices, booking, and availability"1

1

. Launch partners include ResX, CLEAR, BLADE, Aero, SeatGeek, GetYourGuide, Nuitee, and Ten Group, alongside Booking Holdings' own brands like Booking.com, OpenTable, Kayak, Priceline, Agoda, and FareHarbor2

2

. The platform understands natural language, asks follow-up questions, and learns user preferences to make recommendations increasingly personal over time.

Three-Tier Membership Model Targets Loyalty Over Transaction Fees

Lola operates on a membership-based business model with three tiers, marking a departure from traditional travel booking commission structures. The free tier includes personalized search, recommendations, bookings, and promises "no ads, ever"1

1

. The Plus tier costs $5 per month with a 12-month commitment, offering up to 10% discounts on 4- and 5-star hotels worldwide1

1

. Additional Plus benefits arriving this fall include a $300 discount on first Aero private jet bookings, early restaurant table availability through ResX, and discounts on GetYourGuide activities. The VIP tier at $25 per month increases hotel discounts to 15% and adds $150 flat-rate BLADE airport helicopter transfers from New York to JFK and Newark1

1

. VIP members will also receive a $400 discount on first Aero bookings and ResX premium status when these features launch. Both paid tiers include access to human concierge services from Lola. Hafner described the approach as "something like a high-end version of Costco: making money from membership fees, not from the items sold," adding "we're betting on loyalty"1

1

. He expects 95% of users to remain on the free tier initially as the platform proves itself.

Platform Faces Competition from Google Travel, Airbnb, and Meta's Muse

Lola positions itself against experience-focused platforms like Airbnb and Uber, which have expanded beyond core functions into lodging and activities1

1

. The platform also competes with Meta's Muse, which launched earlier with similar AI agent capabilities for booking events, hotels, and flights2

2

. When Muse first appeared, travel booking stocks fell over concerns the AI agent could bypass dedicated travel sites entirely. Lola currently operates only via mobile and desktop web interfaces, with mobile apps and flight bookings on the roadmap. The service faces a significant geographic limitation, operating US-only at launch, which Hafner acknowledged "needs to change"1

1

. Users booking hotels through Lola's Booking.com integration rather than directly with hotel operators may forfeit hotel rewards points and elite-qualifying credit, though Hafner notes that on a $300-per-night hotel room, the 10-15% discount remains significant.

AI Hallucinations Pose Risk to Automated Travel Bookings

Despite the promise of streamlined travel planning, AI agents carry inherent risks through hallucinations and errors. One legal firm documented a case where an AI concierge booked a non-refundable hotel in Naples, Florida instead of Naples, Italy where the traveler was flying2

2

. AI assistants can also generate "confirmed" booking statuses when reservations remain pending. These challenges highlight the gap Lola aims to fill by connecting conversational AI directly to verified availability and pricing systems. Watch how Lola addresses accuracy in real-time booking confirmations and whether its human concierge services effectively catch AI errors before they impact travelers.