‘AI will transform travel UI in the next 3 years’: Rahul Todkar, head of data and AI at Tripadvisor | Technology News

A strategic transformation

In its illustrious 25 years, Tripadvisor has become one of the most recognised brands in travel. Todakar shared that the company has operations in over 40 countries in more than 20 languages, and the platform serves over 300 million monthly active users. It has been hosting over 1 billion reviews and 13 million businesses. Regardless, the company’s recent strategic turn seems to be its most significant transformation yet. 

“Two or three years ago, with our new CEO Matt Goldberg at the helm, we began thinking about how we would shift from a click strategy to an engagement-based strategy,” Todkar said. This shift, according to him, focuses on creating meaningful experiences that encourage users to stay, plan, and engage deeply with the platform rather than simply clicking through to external sites. 

Todkar explained that this new strategy operates on three core pillars: increasing contributions and contributors, expanding trip planners and activity engagement, and driving more direct bookings through their platform. “Our data and AI strategy flows from that,” Todkar emphasised, highlighting how technology serves as the foundation for this transformation.

On AI strategy

Talking about Tripadvisor’s AI strategy, Todkar informed that the company’s approach is unique, adding that AI is not entirely new to their operations. “AI is not new to TripAdvisor. We have done a lot of work in terms of core recommendation models, recommendation systems, personalisation, sort pricing, and a lot of the marketing optimisation,” he said. 

With the advent of generative AI, what has changed, according to Todkar, is largely the consumer expectations. Todkar identified three key shifts driving their AI strategy: the move toward conversational interfaces, multimodal expectations incorporating text, images, and videos, and the demand for hyper-personalisation.

The company’s AI strategy encompasses three broad areas: product and marketing integration, internal productivity enhancement, and external partnerships. “We announced content partnerships with Perplexity, with Alexa, and some of the stuff we’ve launched with OpenAI, Microsoft,” Todkar noted, demonstrating their “Tripadvisor everywhere” philosophy of meeting users where they are.

When asked what he thinks lies ahead for the travel industry, Todkar shared his most striking prediction, which is related to travel interfaces. According to him, the traditional model of date fields, filters, and hotel lists that dominates current travel websites will be replaced by conversational, multimodal experiences driven by social validation and network effects.

Todkar added that this transformation can already be seen with Tripadvisor’s AI trip planner, which was launched 18 months ago. The generative AI-based solution has created over 2.5 million trips, with users showing “three to five times more engagement with us on our site than any other users.” Moreover, the company has reported substantial increases in average revenue per user from these AI-engaged travellers.

Partnership with Snowflake

When asked if he could explain how Tripadvisor’s partnership with Snowflake has helped it to modernise its data and enhance its AI deployments, Todkar said the partnership has been able to deliver real impact. 

“Today, we manage over 10 petabytes of structured data, plus vast amounts of semi-structured and unstructured data – from text reviews (over a billion of them) to half a billion user-uploaded photos. Three to four years ago, all of this lived on-premises. We needed a modern, scalable platform that could handle not just storage and compute but also robust privacy controls and governance. This is when we partnered with Snowflake. Since then, we’ve worked closely together to evolve new features and build solutions that deliver real impact across our business.”

When asked if he could divulge specific benefits observed in terms of performance, cost, and efficiency with such a partnership. “Our storage costs have dropped substantially. More importantly, we’re now 20–30 per cent faster at deploying new solutions on the Snowflake cloud. For example, using the full Snowflake stack alongside tools like Streamlit, we can prototype, monitor, and launch AI models far more rapidly. Overall, this agility has accelerated our product development and go-to-market timelines by roughly 20–30 per cent.”

The travel company’s technology stack demonstrates a pragmatic approach to AI. While Tripadvisor works with multiple providers including Anthropic, Cohere, and Gemini, OpenAI serves as one of the primary partners. “We primarily work with OpenAI because those models have been fantastic,” Todkar explained, citing three critical factors: “cost, latency and accuracy.”