Travel: How OTAs are adding generative AI to test the future of trip planning

The travel industry has been investing in AI and machine learning for several years, with companies like Booking.com, Kayak, and Hopper using the technology to personalise and improve online journeys and predict flight price fluctuations. 

With the arrival of ChatGPT last November, however, travel companies are now considering how the latest iterations of generative AI can further enhance the travel booking experience.  

On the face of it, there are limitations to these kinds of large language models (LLMs), of course. ChatGPT data is mostly limited to before September 2021 – seemingly a big obstacle for travel, whereby information needs to be as up-to-date as possible. In addition, there’s the well-documented issue of hallucinations and half-truths; information provided by generative AI models isn’t always accurate.

However, with the future potential outweighing current concerns, many have recently announced integrations with generative AI applications, with the travel industry leading the way when it comes to investment and adoption of the technology. Here are a few examples, and that the impact it could have on the wider industry. 

Kayak and Expedia launch ChatGPT-powered travel planning tools 

OpenAI launched ChatGPT plugins in March 2023 – it’s latest initiative to enhance and expand the knowledge of its large language model. Companies in the plugin program – two of which include Expedia and Kayak – will provide the LLM with access to fresh data, thereby training the model with more up-to-date and relevant information. Essentially, this means that users can ask ChatGPT for travel information or trip-related suggestions, before being referred to the relevant company for booking options.  

In a blog post announcing its integration, Kayak said that the move “represents a step forward in the world of travel search. By leveraging AI technology to provide more personalized and intuitive search experiences, we’re making it easier than ever for travellers to plan their dream vacations.” 

For Kayak, the aim is also to make interactions between the company and its customers more conversational as well as personal, informing answers with the user’s search criteria and Kayak’s historical travel data. 

As well as its own plug-in, Expedia has gone one step further with the launch of an in-app travel planning experience powered by ChatGPT, which allows the company to meet users inside owned channels. Peter Kern, Vice Chairman and CEO, Expedia Group explained: “By integrating ChatGPT into the Expedia app and combining it with our other AI-based shopping capabilities, like hotel comparison, price tracking for flights and trip collaboration tools, we can now offer travellers an even more intuitive way to build their perfect trip.” In addition to answering open-ended questions, Expedia will automatically save hotels recommended during the conversation.   

The ChatGPT-powered chatbot is still in beta mode, which means app users are currently only offered reviews and suggestions (rather than the ability to book flights or hotels). However, Expedia says that the option to do so will come in time.  

The convergence between conversational AI and travel booking (on travel sites and apps) is where the real opportunity lies in travel, enabling users to both research as well as book, and saving the time and effort involved in manually doing so.  

Rajesh Naidu, senior vice president and chief architect at Expedia Group, summarises this concept. “Imagine if travellers could use it to create a trip itinerary or identify top hotels for their trip and have AI automatically add these recommendations to their Expedia trip board. It can simplify an extremely manual planning process involving sifting through lots of options down to minutes,” he told PhocusWire. 

Customer service and generative AI: a uniquely transformational moment for the industry

Google announces new AI capabilities (and partners with eDreams) 

Google released its rival to ChatGPT, Bard, back in March, following an announcement that it would be opening access to its generative AI tech to developers, enabling them to integrate it into their own platforms. Spanish travel company eDreams was one of the first to do this, stating that it plans to use generative AI to “enhance productivity, streamline development processes, and reimagine customer engagements.” 

While eDreams has previously used its own generative AI models for the personalisation of user journeys, its partnership with Google will enable it to develop a range of products that enable users to interact with AI agents at different stages of their journey.  

The company further explained in a statement: “The AI-powered agents will have a deeper understanding of each customer’s needs, preferences, and context, enabling the business to build the best-tailored travel options and bespoke products for each scenario. This represents a significant leap forward in personalising the travel booking experience while minimising the time and effort required from travellers.” 

It’s likely that Google has plans to insert its AI capabilities elsewhere, too. Google recently released an open-source demo for developers to experiment with and better understand how AI is deployed in travel apps. By combining the PaLM natural language model API with Google Maps APIs, “users can interact naturally and conversationally to tailor travel itineraries to their precise needs,” stated Google. 

This means that, within apps, users can receive recommendations based on factors like the weather, budget, or if they are travelling with children. Google says that through this, developers can “generate a kind of scaled personalization that – when combined with existing customer data – creates huge opportunities in loyalty programs, CRM, customization, booking and so on.” 

Finally, Google has also announced new generative AI updates to Search, including a short answer generated by AI at the top of the results page. At Google I/O, Cathy Edwards, vice president and general manager of search discovery and ecosystems, said that the new capabilities will make Search “smarter and searching simpler… this is really especially helpful when you need to make sense of something complex with multiple angles to explore.” 

A wave of start-ups and new AI projects emerge…

While companies like Kayak and Expedia are integrating generative AI into products and services, several new companies powered by the technology are also emerging. One of the most prominent examples is GuideGeek – created by travel publisher Matador Network – which is a new WhatsApp-based travel assistant built on OpenAI’s ChatGPT technology.  

In addition to the conversational format powered by ChatGPT, GuideGeek also incorporates live flight data from 1,000 OTAs and airline websites, including Skyscanner. Naturally, as with any chatbot powered by generative AI, the results are still likely to be imperfect, but Matador employs teams to monitor conversations and help train the AI.  

Welcome – an app that combines video and maps to help users find and discover businesses – is also focusing on generative AI with a new project, TripNotes. The new travel guide app incorporates data that Welcome has compiled over years to create real-time itineraries for users, including reviews, location and business hours. Speaking to Skift, Matthew Rosenberg, co-founder of Welcome, said that “It’s 100% our focus right now. At least for those of us who come from tech, and from consumer tech, it’s very obvious that this is going to be the way that a lot of things move, but especially things in travel.” 

…while challenges such as fake reviews and legacy systems still need to be tackled

While generative AI is clearly becoming a big focus for the travel industry, there are still challenges to consider. One issue is fake reviews, stemming from generative AI’s ability to create thousands of natural-sounding reviews within seconds. Consequently, the traditional methods typically used to detect fake reviews are likely to be less effective, creating the need for more sophisticated tools such as AI-detection technology, as well as much more intensive investment in social media monitoring from human employees.  

Elsewhere, it’s also been suggested that the hype around generative AI models like ChatGPT could be premature, and that the technology can only be applied to the pre-planning stage – not the far more complex and convoluted booking process. Rathi Murthy, chief technology officer for Expedia Group, recently suggested that the industry has a way to go before the latter becomes a reality.  

“Honestly, the travel industry as a whole has a lot of legacy technology, and it’s not super easy to move in and adopt some of this… We don’t see generative AI, today, being able to replace this whole gamut of services. Planning a trip is just way more complex than just a bunch of questions,” said Murthy. 

However, as demonstrated by Expedia and other investments from within travel, the potential of the technology has clearly compelled some companies to act quickly.