This insight summarises Booking's 5 AI Experiments — and Why Agoda's Is Next by Dennis Schaal for Skift.
A natural experiment in AI strategy. Booking Holdings and Expedia have taken opposite architectural bets on the future of consumer travel AI, and the next two to three earnings cycles will show which one converts.
Booking's five-part portfolio.
- Priceline's Penny — a trip-planning assistant that Booking says is delivering "good CSAT numbers" and "higher conversion" in early data (small sample, but directionally positive).
- Agoda's next launch — the Asia-focused brand is up next with its own AI trip-planner, framed as the most-anticipated of the five.
- Booking.com's flagship AI trip-planner — the anchor experience for the group's largest brand.
- Lola.com — an internal greenfield startup, revived as an AI-first travel product.
- An unnamed West Coast venture — a second stealth internal startup operating independently from the core brands.
Each unit is run separately, letting Booking hedge across brand, geography, and product architecture rather than converging on one design.
Expedia's contrary bet. Expedia has discontinued Romie, its end-to-end chatbot, and pivoted to a coordinated multi-agent architecture — specialised agents that hand off to one another. It also acquired trip-planning startup Layla to accelerate the pivot.
What to watch.
- Whether Priceline's early conversion lift holds as sample sizes grow.
- Which architecture — one unified assistant vs. many specialised agents — wins in retention and repeat use.
- How quickly Agoda's launch closes or widens Booking's lead.