Generative AI is compressing the traditional travel-booking journey, with travellers combining questions about flights, transit visas, and card offers in a single conversation, according to MakeMyTrip Group CTO Sanjay Mohan.

Speaking at Inc42’s inaugural ‘The CTO Summit 2026’, Mohan said this behaviour requires travel platforms to rethink both their search experience and the technology connecting different parts of a booking.

His session, “What Lakhs Of AI Conversations Are Teaching MakeMyTrip About How India Travels”, explored how conversational AI is changing the way users express their travel needs.

A conventional travel app takes users through a sequence: browsing listings, reviewing a selection, entering passenger details, paying, and receiving confirmation. In a conversation, however, users can bring up requirements associated with several of these stages at once.

Mohan cited the example of a traveller asking for a flight with limits on layover duration, no transit visa requirement, and deals on specific bank cards — all within the same request.

For the platform, answering that request means bringing together information that users would otherwise find across separate screens.

Moving Beyond Keywords And Filters

MakeMyTrip is adapting its data systems to handle these overlapping requirements, Mohan said. This includes using knowledge graphs, which connect related information, and search technology that interprets meaning rather than relying solely on exact keyword matches.

For instance, a traveller asking for a “serene getaway” should be able to find suitable properties even if their listings do not contain that phrase.

“If you tell me, ‘I want a vacation in a very serene getaway’. If no one has put that text, then you will not be able to find it,” Mohan said. “I have to understand what you really mean by ‘serene’.”

The shift towards interpreting intent also featured in Flipkart’s session at the Summit, where CPTO Balaji Thiagarajan discussed moving beyond keyword-led search to combine conversational interactions, voice, visual inputs, and content feeds.

For MakeMyTrip, voice is another part of this shift. Mohan said the platform is expanding speech-to-text and text-to-speech capabilities to serve users in smaller cities where typing can be a barrier. Voice interactions already account for a significant activity in Tier III and IV regions, he added.

How Myra Handles Travel Queries

MakeMyTrip uses its in-house AI assistant Myra to handle conversational interactions across customer touchpoints. Mohan argued that the company’s travel expertise helps it address the detailed questions involved in choosing a trip.

“Travel is a very experiential buying thing. You don’t want to look at a toothpaste from five different angles. But if you’re booking a hotel room, you want to know what the view from the window looks like,” he said.

Myra comprises around 10-12 specialist AI agents covering different travel needs, including flights and accommodation. A central coordinating system routes tasks to these agents and brings their responses together.

Balancing Quality, Speed & Cost

Mohan said using the largest AI models for every task increases both costs and response times. MakeMyTrip instead combines frontier models with open-source and open-weight alternatives, depending on the task.

He also emphasised the need for detailed evaluation sets to test whether each specialist agent produces relevant answers.

“A vertical applied AI can be only as good as your eval set. And therefore, for every expert, flights and hotels and holidays, there has to be a very rich set of evals. Otherwise, you will not know whether the answers are relevant,” Mohan said.

These checks cover live prices and availability, the relevance of responses, and whether the system retains details shared earlier in a conversation.

The post AI Is Compressing The Travel-Booking Journey: MakeMyTrip Group CTO appeared first on Inc42 Media.