Zumata harnesses the power of artificial intelligence
When it comes to hotels, there is no dearth of choice in most destinations of the world.
“There are 3,000 hotels in Rome,” Robert Meza, director of business development and strategic partnerships at Zumata Technologies, said.
Somewhere in that number is the perfect hotel for a family with a couple of kids who are looking forward to seeing the Colisseum and would get a kick out of a pizzamaking excursion as well.
The boys are old enough for a room of their own, so their parents would like two adjoining rooms, and a pool would be great.
Mom goes to her favorite online travel agency, one that goes beyond the origin-destination-date formula. She types, “Find a hotel in Rome that is kid-friendly, has adjoining rooms and a pool and is near the fun sights. And a view of the city would be nice.”
In seconds, the OTA narrows the field to a few properties and delivers descriptions and photos.
Since the OTA “knows” the customer, it also suggests the hotels have spa facilities for Mom.
The brains behind the operation belong to Watson, IBM’s cognitive system (and “Jeopardy” champion).
Combing through a vast store of data and images from more than 60 hotel companies and aggregators, Watson can provide Zumata’s customers with very relevant answers to travelers’ questions.
Zumata is a B2B company, chief executive of-ficer Josh Ziegler said.
Its target customers include OTAs, traditional travel agencies, GDSs, loyalty companies and other travel companies.
Travelers don’t know the information is being served up by Watson. “They just think you’re a genius,” Ziegler said.
Zumata uses a combination of Watson APIs, including Alche-myLanguage, Visual Recognition and Retrieve & Rank, to deliver the most relevant results.
Watson, of course, does not know everything.
It even flubbed a few questions when it challenged the top two “Jeopardy” champions in 2011.
So it gauges its own confidence level before answering a question. If it is low — for example, if a traveler would like a restaurant reservation in the hotel’s vicinity — it will refer the query to a human.
It does, however, “learn,” and the Zumata team has “a ton of data that we train it with,” Ziegler said.
“We also use social media to add sentiment and feeling with each interaction.”
Another benefit for Zumata cus-tomers is “dynamic margining.”
If one hotel offers an override of 1%, but you can find a rate from another that offers you a 2% higher margin, you are better off booking the lower rate instead of steering volume to the override.
While Zumata is well-suited to the leisure travel market, the company is also focusing on corporate travel.
In particular, Meza said, it believes it can take a lot of the pain out of using an online corporate booking tool.
“It can take an hour to look for things that are important, like whether a hotel is close to a metro station,” he said.
Zumata is based is Singapore, so it is in the thick of the fastest growing travel market in the world, Ziegler said.
But the greatest demand for its services is in North America, which is why Zumata is teaming up with DHISCO.
Toni Portmann, chief executive officer of DHISCO, notes that it’s an unusual arrangement, given that both are hospitality distribution companies.
But, she said, “taking this in-dustry to the next level and staying competitive requires cooperation and industry alliances.
“That’s why this is so exciting,” she added. “Zumata leads the field in artificial intelligence, which is the future.”
Read article on page 7 >