0
Q2 2016
FIRST
EDITION
01
I S S U E
CHECKING IN
HOW ARTIFICIAL INTELLIGENCE WILL
TRANSFORM THE TRAVEL INDUSTRY
0 3
CAN A NEW K I ND OF COMPUT I NG HE L P
CUS TOMER S MAKE BE T T ER TRAVE L
DEC I S I ONS
0 4
WHY WE NE ED A NEW K I ND OF
COMPUT I NG
0 4
COGN I T I VE COMPUT I NG UNDER S TANDS
THE PROB L EM AND KNOWS HOW TO
F I ND THE ANSWER
0 6
L EARN I NG THE LANGUAGE UN I QUE TO
EACH I NDUS TRY
0 8
A SMOOTHER ROAD TO CUS TOMER
ENGAGEMENT
0 8
COGN I T I VE CONVER S I ON AND
ENGAGEMENT APP L I ED TO TRAVE L
0 9
GA I N DE E P ER CUS TOMER I NS I GHT
1 0
MORE I NT E L L I GENT TRAVE L
DEVE LOPMENT S
1 1
L EARN MORE ABOUT WAYB LAZ ER ’ S
COGN I T I VE TRAVE L SOLUT I ONS
TABLE OF CONTENTS
01 ISSUE | Q2 2016
Can a New Kind of Computing Help Customers Make Better Travel
Decisions?
From the choice of destination to lodging selection and dining, the
travel industry is heavily impacted by recommendations. Before the
Internet, most recommendations came from close friends, or
recognized travel and dining resources, such as travel writers and
guidebooks. For many business and leisure travelers, travel agents
served as both booking agents and as an important source of guidance
in making travel decisions. Today, online booking has in many cases
replaced the travel agent, while online reviewers have become a
major influence in buyer decisions. This environment, in providing
travelers direct access to such a wide array of information and selfservice
booking capabilities, can itself become a source of frustration
and indecision if consumers become overwhelmed by too much data
and too many choices. In addition, online travel booking as it now
stands tends to drive purchasing decisions to the lowest price offered,
which in turn can negatively impact travel industry profits and may
ultimately lower the quality of the traveler’s experience.
But what if the travel industry could influence booking decisions by
helping customers make more knowledgeable choices with a higher
level of confidence? To be effective such a system must be reliable,
robust, and scale to meet demand. And most important, it has to be
intelligent—combining recent advances in areas such as machine
learning (ML) and natural language processing (NLP) to provide the
most relevant travel recommendations.
This white paper explores the development of these advanced
cognitive computing systems and how they differ from the
structured-data systems that came before. We will look at how this
new kind of computing is transforming organizations and businesses
around the world. And finally, we will look at what better travel
recommendations and a more personalized online experience for
travelers could mean to the travel business.
3
Why we need a new kind of computing
Until the recent advent of what is popularly known as artificial
intelligence (AI), the way computers compute hadn’t changed its
underlying structure for nearly 70 years. Since their introduction in
the 1940s, computers had required that data conform to strict rules
and formats. Likewise, the questions we asked them had to be
structured to follow the same well-defined logical path.
These computer algorithms fail with today’s fastest growing and most
pressing information challenges. They struggle understanding big
data and are unable to handle the vast amount of unstructured text
and images found in newspaper and magazine articles, blogs, social
media and elsewhere—unstructured data accounts for 80 percent of
the information humans create and consume today according to IBM
Watson.
We have arrived at a point where a new kind of computing is needed.
We need computers that are able to process information similar to
the human brain.
Cognitive computing understands the problem and knows how to
find the answer
The thing that makes cognitive computing really different is that it
isn’t a pre-programmed answer machine—it’s a learning machine with
the ability to deduce the answer itself. In an article, “Demystifying
artificial intelligence: What business leaders need to know about
cognitive technologies,” (Deloitte University Press, November 4,
2014), authors David Schatsky, Craig Muraskin and Ragu Gurumurthy
look at the business landscape for cognitive computing. They suggest
that while the hype around Artificial Intelligence makes it sound
“more like science fiction than it does an IT investment,” the
technology is real and will be critical to business success.
4
“We distinguish between the field of AI [Artificial Intelligence] and the
technologies that emanate from the field. The popular press portrays
AI as the advent of computers as smart as—or smarter than—humans.
The individual technologies, by contrast, are getting better at
performing specific tasks that only humans used to be able to do. We
call these cognitive technologies… and it is these that business and
public sector leaders should focus their attention on.”
The most famous public demonstration of the cognitive computing
capabilities of AI came in a Jeopardy match in 2011, when the game
show’s all-time biggest winners—Ken Jennings (the most wins at 74)
and Brad Rutter (the most prize money at $3.25 million)—lost to a
new kind of computer, IBM’s Watson. Just to keep things fair, Watson
wasn’t permitted to be connected to the Internet. Like the human
contestants it had to rely strictly on what it had already learned. After
three episodes, IBM’s cognitive computer was the champion with
$77,147 in winnings, compared to Jennings at $24,000 and Rutter at
$21,600.
Since that landmark demonstration, companies big and small have
begun adapting AI in general and cognitive computing in particular to
their business needs. Their goal is to create machines that augment
human intelligence and that can interact with us in a human way.
Augmenting human intelligence is at the heart of this new way of
computing. Up to this point computers provided tools that let us do
things better or faster—they help us with tasks we were already
doing. Cognitive computing, on the other hand, can help us do things
we hadn’t thought of doing before.
5
The ability for a computer system to understand natural language—the
unstructured written information that accounts for the bulk of
information produced and consumed by humans—is a significant
breakthrough. For the first time, machines can comprehend and extract
semantic knowledge from any unstructured data resource, including:
Research papers
Technical, business and news articles
Books and essays
Personal blog articles, reviews, tweets and social media posts
Images and photos
To interpret information like a person, the cognitive computer breaks
down each sentence into its syntactic structure. It looks at the
relationships of words and phrases and their context within the whole
document. It then looks for relationships with other words and phrases
from other documents available to it. This ability to absorb knowledge
like a human helps the computer answer human questions, posed in
natural language, with truly human-sounding answers. The objective is to
understand the intent of the user’s question and use that understanding
to develop logical responses, draw inferences, generate potential
answers, and settle on the best response backed up with source
information.
Learning the language unique to each industry
When someone begins working in a new industry, the first job is to learn
the language specific to that industry. In specialized fields, practitioners
even have to learn the thought patterns that are unique to it. In medicine,
for example, there are multiple domains, each with its specialized
language and knowledge base.
6
When IBM Watson was enlisted in the battle against cancer, it needed to
learn about the many different types of cancer, their symptoms and
treatments. It also had to learn to distinguish various cancers from other
diseases with similar symptoms. Before suggesting a course of treatment,
Watson had to be able to take into account side effects that can vary
from one patient to another.
All this learning requires the guidance of human experts. They are key in
developing a corpus of knowledge. Assembling the corpora starts with
digesting masses of information from every available and relevant
source. The real building then begins with domain experts assisting in
evaluating the information and discarding any items that are out of date,
irrelevant to the problems that are to be addressed, or held in poor
regard by the domain’s peers.
Next comes the training phase, in which the computer learns how to
interpret information in a process known as machine learning. With
supervised cognitive systems like Watson, experts will upload training
data in the form of questions and answers. The purpose isn’t for the
system to memorize answers but to uncover the underlying patterns of
thought that lead to answers. These patterns tend to be unique to a
particular domain or industry.
The system continues to learn once it begins interacting with end users.
There are regular updates of information and periodic reviews by
experts. Over time, cognitive computing systems grow more capable of
responding to complex situations. Users receive quick answers with an
array of recommended choices backed by evidence. They even get help
uncovering new insights and patterns that had been hidden in the mass of
information the cognitive system has been fed.
7
A smoother road to customer engagement
Rob High, an IBM Fellow who is vice president and chief technology officer
for IBM Watson, describes in a recent blog article how Internet searchers
too often end up going down a tortuous path when trying to engage with a
company. As High sees it, “the onus is on you [the customer] to discover your
own answers, and that experience can be very disengaging.”
It’s a simple matter now for advertisers to use a browser’s search process to
discover specific interests and target their advertising accordingly. But the
path from detecting interest to consummating the sale is littered with
failures to engage. At the very point when customers want answers or need
assurance from an experienced advisor, they end up lost on the company’s
website. Or they go wading through hundreds of reviews looking for
someone whose opinion they can trust. Or worse, they get caught in 800number
purgatory, pressing phone keys in the hope of getting a real person
on the line.
High sees customer engagement as one of the next big areas where AI and
cognitive computing will make a difference. Whether it’s financial products,
insurance, telecommunications or travel, this new type of computing can
help potential buyers in their quest to know more about a company’s goods
and services.
Cognitive conversion and engagement applied to travel
It is now possible to take what we’ve learned about cognitive computing to
deliver fast, accurate, personal recommendations and apply it to the travel
industry. With these new intelligent engagement capabilities, customers can
now interact with hotel or travel company websites and apps in more
meaningful ways.
8
Travel marketers can use this technology to help travelers fashion a more
personalized travel plan and allow them to book with greater confidence. By
combining NLP and ML, travelers can request specific trip characteristics
and be presented with a set of results that best match. For example, showing
insights and images that are relevant to families when searching for a room
with 2 adults/2 children. Providing increased accuracy in the results leads to
the increase in likelihood of the traveler completing that booking; improving
both conversion and brand loyalty.
This personalized conversion and engagement capability can be extended
through the traveler’s entire journey. From pre-arrival communications to an
in-destination app or kiosk, travelers can experience any destination like a
local. As customers engage and use the technology, their experience and
commitment to a brand grows deeper. For travel marketers, that means:
Increases in direct bookings
More website stickiness / engagement
More social shares
Greater customer trust
Higher perceived value
Increased brand loyalty
Gain deeper customer insight
In addition to enhancing customer interaction, companies can also capture
the buyer-focused intelligence they need to make better marketing and
operational decisions. By integrating cognitive computing into their CRM,
marketers can reap deep insights from their customers’ declared, observed,
and inferred inputs.
For example, if a hotel notices an increase in “Where can I find live music?”
queries, they can target those specific customers with music related
marketing and messaging. Discovering these sorts of opportunities to
promote targeted personalization increases the likelihood of repeat business
and long-term loyalty.
9
More intelligent travel developments
The services being developed for AI’s cognitive computing platforms are
growing at an exponential rate. Following are some services currently
available that can be applied to the travel industry.
Natural Language Classifier (NLC): This service applies cognitive computing
techniques to predict the best matching, predefined class for any given input
query. Depending on the output class, the cognitive travel platform is able to
return more accurate results, or recognize if the query lacks specific
relevant information. For example, if a user searches for “Where can I crash
in Austin tonight,” the NLC understands that the query is actually asking for
suggestions on places to stay or lodging options in Austin—not a car crash
—even though it wasn’t explicitly specified in the question.
Relationship Extraction: To help analytics engines more easily understand
the meaning of a sentence, this service first extracts all entities that were
mentioned in the sentence, followed by the specific relationships between
those entities. Based on a user’s blogs, emails or tweets, the cognitive travel
engine is able to identify all the places and events that he or she wrote about
and thus provide more personalized recommendations in the future. For
example, a user’s blog can easily be analyzed to pull out key themes. If every
time the user mentions lodging in the blog, the focus is on small boutique
hotels, a travel company would know to show boutique hotels at the top of
the search results for this user.
Concept Expansion: This service analyzes a body of text for words or
phrases that are contextually related. For example, it knows that “The Big
Apple” refers to New York City, “honeymoon” indicates a romantic type trip
or that “wanting to go off the beaten path” is asking for a unique getaway
travel experience.
10
Personality Insights: Understanding a user can be enhanced by looking at that
individual's posts, tweets, or online reviews. This service uses that information
to identify a user’s intent and behavioral traits. These insights in turn can lead
to showing the most relevant content to that specific user. Users could connect
via social media and the travel company could begin to create a persona unique
to each one of them. For example, for an outdoors person (visible through their
hiking, biking and fishing photos and content), the travel company could
suggest outdoor style accommodations or packages.
Learn more about WayBlazer’s cognitive travel solutions
WayBlazer is a cognitive travel company leveraging Artificial Intelligence and
IBM Watson to help every traveler find the perfect hotel. This is accomplished
by analyzing cues and triggers from the traveler and personalizing the search
results.
Beyond just finding the best hotels, the cognitive technology dynamically
merchandises the images and reviews to present the hotel in the most relevant
way for that particular traveler. The combination of the best hotels and the
personalized presentation and content leads to increased traveler engagement
and improved conversion rates. WayBlazer delivers this technology to travel
companies already selling hotels and to travel related companies that want to
add additional revenue streams from hotel bookings via API or rapidly
deployable, hosted platform.
11
Author:
Felix Laboy, CEO
Felix was formerly President & GM at Sabre Hospitality Solutions, CEO & CoFounder
at E-site Marketing, Executive VP at the Puerto Rico Convention Bureau,
and Executive positions at Ritz-Carlton, Westin, and Four Seasons.
Felix is a Cornell University School of Hotel Administration graduate and member
of the Stanford University Parent’s Advisory Board. He also enjoys participating in
triathlons and has traveled to all seven continents including Antarctica and a
recent climb to the top of Mt. Kilimanjaro. His favorite travel destination is Madrid.
Contributing Experts:
Ricky Stillwell, Lead Engineer
At WayBlazer, Ricky focuses on the intersection of data and technology, having
developed the initial graph, NLP, and backend frameworks. Most recently, he was
a member on the IBM Watson core development team and helped mature
Watson from a Jeopardy competitor to a doctor and research analyst. Previous to
Watson, he was on IBM's emerging technology team which focused on bringing
new technology to market in a fast paced development environment.
Ricky holds a B.S. in Computer Science from Texas Tech University. His
professional interest lies in applying new technology to markets that have not yet
been transformed. He likes to spend his free time restoring vintage computers
from the 1970's.
Melanie Tosik, Research Engineer
Bringing her linguistic expertise to WayBlazer, Melanie leads a research team
focused on developing complex, real-world applications of NLP and ML.
Melanie holds a B.S. in Computational Linguistics from the University of Potsdam.
Her main research interest is in computational semantics — the computation and
representation of meaning in natural language.
WayBlazer
Austin, Texas
howdy@wayblazer.com
12
1
Publizr
Index
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
Home
