TL;DR
Expedia Group is transitioning from efficiency-focused AI to agentic systems that anticipate traveller intent and execute actions autonomously. Smart Trip AI, announced in 2024, merges first-party data with partner insights to generate personalised, mood-driven itineraries. The platform combines 70 petabytes of historical travel data with third-party context (weather, local events) to create self-healing trips that proactively monitor conditions and suggest rebookings before travellers identify issues. Expedia is embedding travel services across OpenAI’s Microapps and Operator, and Microsoft Copilot Actions, enabling chat-to-booking workflows. Over half of travel executives are experimenting with agentic AI according to Skift Research and McKinsey.
Opening
As the travel industry approaches the Phocuswright conference, artificial intelligence deployment is shifting from operational efficiency to emotional anticipation. Expedia Group’s Chief Product and Technology Officer for B2B, Karen Bolda, characterises the evolution: “You don’t dream about making a booking. You dream about moments.”
Context: Three-Layer Data Strategy
Expedia’s agentic AI strategy combines three data layers to fuel predictive personalisation. First-party data comprises the company’s global travel dataset. Zero-party data captures what travellers willingly share about preferences and intent. Third-party data provides real-time context including weather patterns and local events.
This combination enables systems to infer trip type—mission-critical versus casual getaway—and tailor recommendations and rebooking thresholds accordingly. Bolda notes: “The more we start combining everything we know about you as a traveller and what’s going on in the place that you’re going to, I think it gets better and better.”
Smart Trip AI generates itineraries from natural language prompts like “walkable weekend near lavender fields,” merging mood-driven discovery with structured booking workflows. Upcoming iterations will use third-party data alongside historical patterns to create self-healing trips—monitoring inbound flights and weather conditions to suggest rebookings before travellers recognise disruptions.
The platform is embedding travel content and services across partner AI assistants. Integration with OpenAI’s Microapps and Operator, plus Microsoft Copilot Actions, enables travellers to move from simple chat prompts (“show me weekend options in Santa Fe”) to complete itineraries and bookings within minutes.
B2B Infrastructure as Distribution Layer
Expedia’s B2B platform powers 70,000 businesses and 160,000 travel agents globally—from banks and airlines to retailers—functioning as an AI learning and distribution accelerator. Bolda has established a B2B Labs team experimenting with emerging AI models alongside partners: “We’re co-developing with our partners and figuring out the most impactful ways we can use AI to remove friction and improve traveller experiences.”
The company is rebuilding infrastructure to be model-agnostic, allowing partners to integrate proprietary assistants without constant rewrites. This positions Expedia as an AI travel operating system rather than solely a booking platform, enabling third parties to build upon its services layer.
Looking Forward
By 2026, Bolda predicts predictive, personalised assistance will transition from experimental to standard. The combination of historical data depth, real-time context integration, and autonomous action capability could shift competitive dynamics from price transparency toward anticipatory service quality.
The strategy’s effectiveness depends on whether correctly anticipating traveller needs creates sufficient emotional switching costs to counteract loyalty programme portability and price comparison ease. For consumers, the trade-off involves convenience gains against data sharing requirements and potential platform lock-in as AI systems become more personalised and effective over time.
Source: Forbes