[ad_1]
When Kim Bennett founded her travel company, AtlasGuru, the idea was that travelers could plan vacations themselves with the help of recommendations from a community of well-versed travelers. It was 2019.
By 2023, she realized that users didn’t need to physically sift through hundreds of long-form articles and curate content themselves. Artificial intelligence could do it for them. Last August, Bennett launched her own generative AI product that forms highly customized travel itineraries, complete with maps and links, using a variety of variables and her encyclopedic body of content.
“You create a one-of-kind itinerary in seconds that’s tailored to your specific interests, and it leverages our crowdsourced itineraries to power the AI-generated itineraries, even providing links to these related itineraries,” says Bennett, who serves as AtlasGuru’s chief executive officer. “This automation gives travelers the speed and magic-like personalization of generative AI, with the human touch that instills confidence in users that their trip will be amazing.”
AtlasGuru is just one of several travel companies looking at generative AI to assist in the travel planning and booking process. For AtlasGuru, it’s about rethinking user-generated content from long-form reviews written by members of its community. Travel booking services like Expedia and Tripadvisor have launched their own itinerary-building products based on reviews and hotel pages, and Virtuoso, a network of luxury travel advisors, is also leaning on AI to help them personalize luxury vacation experiences.
Bennett’s team is realistic about the current state of AI: No one is going to book an international trip based on the output of an algorithm alone. But what it does is serve as a highly tailored starting point, and that’s how all these companies view generative AI.
“AI supercharges the travel advising profession by acting as an incredibly effective assistant, allowing the advisor to focus on the elements of their profession that are uniquely human: service and trust,” says Matthew D. Upchurch, chairman and chief executive officer of Virtuoso. “You want to have someone you trust behind it.”
Whether a prospective traveler is working with a travel advisor or scrolling through travel websites to plan a trip themselves, AI is becoming an increasingly useful tool. The most obvious is assistance from generative AI, like at AtlasGuru. A user will input their desires for a trip, and the tool will produce a customized itinerary, including hotel information from a travel company’s proprietary database. For Tripadvisor, it pulls data from over 8 million businesses and more than 1 billion reviews and opinions; Expedia has over two decades’ worth of data on travel behavior and preferences, which amounts to quadrillions of combinations.
Tripadvisor’s chief product officer, Sanjay Raman, explains that the company’s massive number of reviews is great for marketing, but understands what it feels like to be presented with that as an independent traveler. “It can be daunting to find what’s relevant to help make decisions,” he says.
Tripadvisor launched an itinerary builder in mid-2023 using ChatGPT’s large language model, which gives travelers an easily digestible itinerary complete with links to restaurants, hotels, and things to do from Tripadvisor’s platform. The product’s ongoing goal is to dive deeper, surfacing hidden games and incorporating more conversational aspects, but much of that development will come with time. Raman says that members who build the itineraries generate threefold higher revenue than the average Tripadvisor member.
Both Tripadvisor and Expedia employ AI for hotel recommendations as well to share amenity details and specific attributes of a property, such as “breakfast included,” Wi-Fi quality, free parking, or lenient cancellation policies. Expedia also leverages natural-language processing to enable its self-service virtual agent to resolve traveler queries quickly. To date, Expedia has powered more than 30 million virtual conversations with the tool.
On the B2B side, hospitality companies are using AI to improve high-touch service by analyzing massive amounts of customer relationship management data and providing detailed insights into behaviors and preferences. Natural-language processing examines customer feedback and personalizes offerings. Robotic process automation streamlines routine tasks, such as booking-information and data entry, to free human energy for a focus on customer service. Predictive analytics anticipate future booking patterns and help with inventory management and the customization of marketing efforts.
Upchurch explains that this enables a more nuanced understanding of desires during planning and booking. It supports the “concept of the connected trip” by allowing pertinent information to pass along to everyone who touches an individual’s travel experience. And it helps with repeat business as properties, advisors, and other travel companies can fine-tune recommendations and experiences to each client’s distinct preferences.
Raman has also seen the impact of AI on the operations of hotels by providing informative transparency into the minds of visitors. He says that hotel owners and operators are using Tripadvisor’s product to see real reviews of their properties on the third-party site, and value the results because they know the source, Tripadvisor, is a trusted entity.
Today, the best AI applications for travel still incorporate a human touch. Bennett notes that people are already skeptical of complex algorithms on autopilot, like self-driving cars. And as a business owner, she recognizes the faults of generative AI, with hallucinations that, at times, can confidently give incorrect answers.
While generative AI works out its kinks, she adds, human-created content enables validation. Upchurch says that skeptics in the travel industry are quickly being won over by the value AI adds when paired with human expertise to improve their processes and build their businesses.
As Raman reiterates, “People-powered perspectives build trust.”
[ad_2]
Source link
