Google search engine and AI – A masterclass for travel agents from Wishy Arora

Google search has evolved rapidly as user interactions with technology change, travel industry represetnatives have been told in Berlin. Today’s users are submitting queries that average two to three times the length of previous search queries. They are seeking experiences that incorporate context, personalisation and conversation.

Wishy Arora told delegates at ITB 2026 in Berlin that AI mode processes these queries by breaking them down and combining information from multiple sources to create responses. Sources include the web, weather data, partner data, feeds, maps and geographical information.

Mr Arora explained that users once submitted queries of three to five words on travel topics such as things to do in Nashville or flights to Berlin. Search engines answered those queries and users combined the answers to form plans and handle bookings. This agreement no longer applies as users now submit queries with details on locations, companions, interests and preferences.

He said answers draw on large language models, a knowledge graph that holds facts about the world, content in forms such as videos, reviews, audio and photos, price and availability information, maps, personalised details from users who opt in and agents that perform planning and booking tasks. Gemini serves as the large language model in this setup.

Agents complete tasks that require time, move toward proactive support with context that persists across devices so experiences continue on different surfaces. Flight searches are coming with flexible dates and destinations with data on pricing to show options such as travel to Palma at a discount of 32pc in May. Hotel queries are arriving with details on proximity to trails, heated pools and childcare services. Planning for trips by train or car is incorporating transit data to suggest sequences of cities and routes that include the Bernina Express between Switzerland and Italy, or details on live music start times in pubs from reviews.

Bookings can now occur inside AI mode with selection of dates, rooms and rates followed by confirmation. Partners serve as the merchant of record and provide support after booking. The combination of large language models with data from partnerships enables trip planning and booking.

In his masterclass, Mr Arora told delegates that travel companies should build web content that follows search engine optimisation practices and includes structured data based on schema.org for hotels, locations and attractions. They should consider feeds supply data on lodging, attractions and travel options that contain images, titles, descriptions, property attributes and room information with prices that match those on company websites. Loyalty programmes can be linked through protocols such as OAuth and Google wallet to surface benefits and targeted rates. Companies can implement payment systems that reduce friction in checkout and provide performance data. Companies should focus on operations and customer experiences that lead to reviews.

Wishy Arora shared: “Users are giving us all of this context and in return they expect a different response. They expect AI to do something for them and they know they are learning that it can do that for them. We like to say that context unlocks magic and that’s what we really want users to know.”

“Users are using LLMs very differently than they used to use search. Queries are much longer, much more context. On average, we are seeing two to three times longer queries in AI mode and Gemini than we used to see on search.”

“At the end of the day the best thing you can do is focus on your business. Focus on giving users the best experience possible experiences that they will enjoy and they will actually share and talk about and that is the best way to get your brand surfaced.”