And why not? There’s little to do online that’s more annoying than booking flights, hotels and everything else you need just to take a break, or be somewhere for work.
Mezi’s newly designed app allows users to find and book flights and hotels and make restaurant reservations wherever they want to travel by simply typing a bit about what they’re looking for into a messaging-style interface, or chatbot.
Rather than relying on AI from the likes of IBM Watson or other providers, Mezi has built its own deep learning and natural language processing systems to make its app able to converse with users, and to filter out three options that meet their needs each time they make a request.
Travel apps, from big players like Kayak and Google to growing startups like Hitlist or Skyscanner, already help users plan trips or find great deals on flights, hotels and more. But most require users to link to additional sites to complete their purchases.
Mezi handles not just “discovery,” for its users, but also transactions. Users only have to set up their payments and personal info one time. And each time they ask Mezi to book them a trip, they have to confirm the whole itinerary before booking and purchases are complete.
According to CEO Swapnil Shinde, the average travel order placed through Mezi in its 10 months of existence as an all-purpose, personal assistant app was $505.
Travel requests grew three times faster than personal shopping, retail-related requests on Mezi, which had a much lower average dollar amount spent per order, he added.
That’s what fueled the company’s decision to specialize around travel.
Mezi is presenting at the PhocusWright travel industry conference this afternoon. The CEO will be speaking about how artificial intelligence is poised to transform online travel agencies, airlines, hotels and even local activities businesses.
Shinde said his startup aims to help all of the existing players, whether they’re airlines, hotels or “OTAs” (online travel agencies) begin to benefit immediately from AI.
“What AI does,” he said, “is to allow travel providers to give customers a concierge or white glove service, instead of telling them to DIY online.”
Because Mezi stores information from each interaction with its users, the app learns a travelers’ preferences and requirements over time, and applies this knowledge to future conversations and to the decisions it makes about which itineraries will prove most appealing for the user.