personal assistant that lets users control its applications
by voice, via smartphones or tablets.
IFS lab director Bas De Vos said the prototype app
only does speech recognition now, but he wants to add
natural language processing to interpret questions and
orders. De Vos predicts it may take three to five years
to do that. Hurdles include dealing with accents and
background noise, ensuring reliability and making a
solid business case for investment.
New tools are beginning to spread outside hangars. Delta call centers use speech recognition, and
China Eastern Airlines lets passengers use a variant of
Microsoft’s Cortana to get flight updates and boarding
notifications, pre-order meals and contact crew, other
passengers or family members on the ground.
Vice president of travel and transportation Raimon
Christiani said IBM is developing use cases for smart
machines in several areas.
First are boarding-pass kiosks that accept only
passports and punched-in data now. Combining
speech recognition with natural language processing
will enable kiosks to respond more like human agents
to requests and questions, for example about baggage,
flights and options. Te same techniques could enable
call centers to handle recurring questions and gather
feedback, saving staff for more complex interactions
For passengers, IBM offers cognitive travel advisor to assist in travel shopping and travel planner and
travel track for corporate travelers. Both applications
would be similar to Apple’s Siri, but smarter about their
specific advisory duties, like chatbots for travel.
IBM’s maintenance tool is Airline Cognitive Line
Maintenance, which uses natural language processing
to search maintenance logs for part histories and repairs
and is already being used in Asia.
Similarly, natural language processing could help
airline staff navigate and understand the highly complex rules and regulations affecting fares and ticketing
procedures, customs rules for cargo, interline agreements and airline accounting.
And smart machines could speed up onboarding of
airline staff through computer-assisted education and
training. Christiani calls this application airline expert
builder and said it would be like assigning a personal
mentor to new airline employees.
IBM can handle speech recognition in English,
French, German, Japanese and Italian and is work-
ing on Arabic. Watson and IBM’s familiarity with the
airline business make implementation easier. Christiani
predicts some of his applications will begin to be used
Kevin O’Sullivan of SITA Lab notes the new tech-
nologies cover a lot of territory. He too expects kiosks
will listen to and understand customers, but “we’re
still in the early days.” For example, handling different
English accents is still a challenge.
Eventually, O’Sullivan expects intelligent machines
to be used for many sorts of airline duties, for instance
helping with handling all the variables and choices
involved in disruptions. Machines may someday sit in
airline conference rooms and respond to tough disrup-
tion questions, a bit like Amazon Echo has become a
useful member of the O’Sullivan family kitchen.
SITA Lab continues to look at wearable technologies
to exploit intelligent assistants for maintenance, and
O’Sullivan expects the next generation of wearables like
Hololens to make this practical. Generally, he thinks it
will take another year or two for the promise of smarter
machines to reach airline operations. “Tere is still a lot
of grunt effort to make these technologies work. You
have to spend time understanding how to feed data in
and how to get answers out.”
Henry Harteveldt, co-founder of Atmosphere
Research, sees “enormous potential” in smart technolo-
gies. Travel apps like Lola and Hyper have already
joined Siri and Cortana in transforming consumer
experiences. Harteveldt also sees applications within
airlines, for example in handling scheduling for pilots
Harteveldt notes one additional advantage of using
speech tools in functions like maintenance, part-order-
ing, fault reporting, call centers and baggage handling:
an accurate record is kept for later review or big data
Moreover, “people are using smart machines in their
personal lives and will expect to do the same in their
work lives.” Airlines will not likely be the first
businesses to adopt the new tools, but should follow
the leaders quickly.
“You have to spend time understanding how to feed data in and
how to get answers out.”—Kevin O’Sullivan, SITA