beyond bots and robots - ibm · cognitive robots ibm watson’s robot that is powered by...
TRANSCRIPT
1
Beyond bots and Robots: Exploring the unrealized potential of cognitive
computing in the travel industry
Steve Peterson, Global Travel & Transportation, Industry Research Leader,
IBM Institute for Business Value
2
� Why are airlines interested in
cognitive computing?
� What progress has been made
on cognitive transformations?
� How can airlines get the most
from cognitive computing?
Cognitive
Transformation
Actions
1
2
3
1
2
3
3
Advanced analytics and cognitive are two very different animals
Advanced analytics:
• Traditional rules-driven
systems
• Organize and evaluate
structured information
• Apply mathematical rules
to the amassed data
• Typically produce
forward-looking
projections, or predicted
trends, on the basis of
prior period results
Cognitive computing:
• Next-generation
information systems
designed to accelerate,
enhance, and scale human
expertise
• Systems that continually
build knowledge and
learning
• Typically understanding
natural language and are
designed to interact more
naturally with humans
4
To look beyond headlines, we spoke to leaders with real cognitive experience
Survey respondents by region
Europe Americas
Asia Rest of World
302responses
Survey respondents by role
CEO / Board CDO
CHRO CIO / CTO
CMO SVP / VP / Director
302responses
5
Leaders apply the insights they derive from real cognitive analysis
Level of sophistication applied to analytic analysis
De
gre
e t
o w
hic
h in
sig
hts
are
ap
pli
ed
to
b
usi
ne
ss d
eci
sio
ns
5%38%
7%
1%8%17%
5%
14%
6% Leaders – have developed and applied robust cognitive capabilities
Learners – have expertise in advanced analytics, but not cognitive computing
Laggards – have some exposure to analytics, but only nascent capabilities
Survey respondents by degree or analytic sophistication and level of
application of analytic insights
6
� Why are airlines interested in
cognitive computing?
� What progress has been made
on cognitive transformations?
� How can airlines get the most
from cognitive computing?
Cognitive
Transformation
Actions
1
3
1
2
3
2
7
For most, the transformational journey to a cognitive future is just beginning
6%
56%
33%
3%
24%
58%
17%
1%
0%
10%
20%
30%
40%
50%
60%
70%
None 1-3 projects 4-9 projects 10 or more
Leaders Everyone but Leaders
Number of data analytics projects that will have truly cognitive
elements in the next two years
8
So far, the travel industry has focused its cognitive investments on bots and robots
Bots
� Seen by many as a low risk opportunity to test cognitive technologies
� Supporting select customer service interactions across all travel segments
� Leveraging a mix of machine learning, artificial intelligence and natural language processing systems
Robots
� Although less common that chat bots,
robots are emerging to interact with
travelers in service settings
� Initial efforts have focused on informing and
entertaining travelers, robots will soon be
equipped with sensors to collect data on
travelers and their environment
9
Alterra’s Marina helps travelers book flights
and hotels using messaging bots via
Facebook Messenger, Skype, Slack,
Telegram, Kik, WhatsApp, WeChat, etc.
• Recommends trip destinations
• Recommends trip activities
SnapTravel helps travelers book hotels using
messaging bots via Facebook Messenger,
Slack and SMS; if an user stumps the bot
with a unique request, a human agent will
intervene and assist the bot
• Provides specially negotiated hotel rates
Bots were a good place to start, and have been well received by customers
Many cognitive travel startups are using bots to provide concierge type services
An artificially intelligent service now
available at twelve Edwardian Hotels-owned
Radisson Blu locations that is able to report
on hotel amenities, give directions and tips,
and receive guest complaints in a matter of
seconds via SMS
10
Indeed, the travel industry is increasingly dependent on bots for customer service
Kayak’s Slack bot helps
travelers with flight, hotel,
and rental car searches via
queries (i.e. “/kayak flights
from newark to paris france
April 4 to April 11.”)
• Tracks flight status
Cheapflights’ Facebook
Messenger bot helps travelers
book flights and hotels
• Provides travelers with the
average cheapest round-trip
flight in the future
• Recommends trip
destinations
Expedia’s Facebook
Messenger bot helps
travelers book hotels
• Provides top 5 hotels and
prices for selected dates
• Confirms booking
• Provides itinerary
information
Skyscanner Facebook
Messenger bot helps
travelers book flights
• Yields the cheapest
flights
• Recommends trip
destinations
Most travel bots have adopted only mild cognitive capabilities, but future iterations
of these systems have the potential to further improve travel experiences
11
Cognitive enabled interactive robots are also finding a role in service delivery
Robot created by BlueBotics that scans
passengers’ boarding passes, guides them
to their gates, and lets them know how
long until they reach their gate
Robotic kiosk created by VGo in collaboration with
the Verizon Innovation Program that answers
passengers’ questions (i.e. where to pick up tickets,
boarding times) and assist customers with any
incidentsRobot created by SoftBank that answers
passengers’ questions in Japanese, English, and
Chinese about airport facilities, important flight
information, and weather forecasts
JAL’s “NAO” at Tokyo’s
Haneda AirportKLM’s “Spencer” at
Amsterdam’s Schiphol
Airport
JetBlue’s “Crewbot” at
New York’s JFK Airport
12
The hotel industry has been particularly eager to deploy cognitive robots
IBM Watson’s robot that is powered by
WayBlazer assists with visitor requests,
personalizes the guest experience and
empowers travelers with more information to
help them plan their trip
Savioke’s robot that autonomously relays
items (i.e. extra towels) between hotel staff
and guests via SMS or phone call request.
Four robots are currently deployed in
California
MFG Automation’s robot that helps guests with
their luggage storage, carrying up to 500 pounds
“Yobot” at Yotel Hotel in
Midtown Manhattan
“Connie” at Hilton
McLean in Virginia
“Relay” at numerous
hotels in California
13
But bots and robots under-represent the true potential of cognitive computing
Bots and Robots
What they don’t yet do
� Leverage multiple complex systems to meaningfully engage travelers
� Take advantage of the power of cognitive computing to identify relationships between multiple datasets
� Meaningfully impact the core travel business
What they do well
� Frees up resources to focus on other, more difficult customer service issues
� Helps get the traveling public comfortable with natural language processing
� Harvest the information within a few discrete domains
14
Two types of transformations will emerge as cognitive reaches its full potential
Customer Personalization ProgramsOperational Improvement Programs
� Prior industry efforts to optimize
operations have focused on single silos
� Large operational improvements depend
on connections between silos, which is
where cognitive can excel
� Personalization has long been on the travel
agenda, but progress has been slow
� Cognitive will be used to scale understanding
of traveler patterns and preferences
� Cognitive puts real personalization within
reach
15
Travel companies expect cognitive to improve operational efficiency
Operational Improvement Programs
53% of leaders, and 52%
of everyone else said they
were planning to use cognitive
computing to improve operational
performance
Types of Projects
� Comprehensive cognitive
maintenance programs
� Flight planning assistance
using cognitive tools
� Dynamic employee and
asset re-scheduling during
operational disruptions
16
Cognitive data discovery gives MRO a tool for diagnostics and decision support
Leading Asian airline innovates
around legacy maintenance
challenges with a cognitive-enabled
app
Cognitive computing amplifies human expertise to uncover important diagnostic clues in unstructured data.
Business challenge
Aircraft maintenance is a complex puzzle involving hundreds of subassemblies and thousands of parts. Technicians face immense pressure to solve the puzzle quickly to avoid flight delays and cancellations. Human
intuition only goes so far in synthesizing years of unstructured data. The airline needed a way to unlock vast amounts of information in its maintenance records, enabling aircraft technicians to diagnose and solve problems quickly.
Cognitive transformation
The airline uses a cognitive discovery system to analyze structured and unstructured data from its ERP environment, detecting patterns in how and why equipment fails to diagnose and fix problems. The solution uses natural
language processing and advanced content analytics to interpret and cross-reference text, finding connections that humans may miss, such as the effect of changing seasons on tire replacements or wear and tear on seats.
17
Two types of transformations will emerge as cognitive reaches its full potential
Customer Personalization Programs
Types of Projects
� Traveler recognition
programs across
touchpoints
� Personalized pricing and
offers
� Loyalty personalization
(e.g., customized incentives)
17% of leaders, and 29%
of everyone else said they
were planning to use cognitive
computing to improve travel
personalization
18
Hotels illustrate the potential of cognitive by delivering personalized insights
“Cognitive search will not only find you the perfect hotel but present it in the perfect manner.”
—Scott Goldberg, Vice President, WayBlazer
Business challenge
Many business-to-business (B2B) companies in travel are focused on driving traffic to a hotel’s website but are not as focused on closing the deal. To improve the number of bookings from site users, websites must make it easier
for users to find the options they want. The founders of WayBlazer wanted a way to understand all the unstructured social, review, image and other data that travelers were seeking.
Cognitive transformation
WayBlazer developed an online travel application and application programming interfaces (APIs) that do more than just return a list of travel options for basic parameters. They use cognitive computing to interpret
unstructured social sentiment, images and reviews across thousands of sites, then maximize conversion and revenue by delivering specific answers and recommendations for each traveler.
Answering natural-language
questions with unprecedented
personalized insight and advice
19
� Why are airlines interested in
cognitive computing?
� What progress has been made
on cognitive transformations?
� How can airlines get the most
from cognitive computing?
Cognitive
Transformation
Actions
1
3
1
2
3
2
20
Airline executives must set the right cognitive course and execute effectively
Cognitive Plan Cognitive Actions
• Where is my cognitive strategy?
• What are my top cognitive priorities?
• What are the milestones in my
cognitive roadmap?
• How can my enterprise maximize
return on data?
• How can our apps support our cognitive ambitions?
• How can we facilitate continuous cognitive enterprise learning?
21
Deciding where to invest is not as easy as it was when cost was the clear priority
Historical Language
� Fleet efficiency
� Asset utilization
� Customer retention
� Customer Self-service
� Process standardization
� Scale
Consistent Differentiated
Operations Experiences
Co
nta
inG
row
Co
sts
Re
ve
nu
e
Dominant Strategy 2003-2013
Dominant Strategy 2013-2023
Future Language
� Personalized
experiences
� Journey management
� Value to customer
� Empowered employees
� Lifetime loyalty
� Scope
Differentiation strategies demand innovation
22
Cognitive transformation demands clear alignment of objectives and investments
implement cognitive
solutions as quickly as
possible to drive
competitive advantage
ADAPT
EXPLORE
PIONEER
ADOPT
adopt cognitive solutions
only after they have been
proven to limit risk
acquire and integrate
available solutions to well
defined business challenges
develop and customize
cognitive solutions to
satisfy unique needs and
objectives
Cognitive
adoption
Investment
approach
23
Cognitive transformation requires strong execution with data, apps and projects
Deliver actionable insights Drive program excellenceDelight in data complexity
Data Apps
Execute cognitive programs,
not projects
Prioritize business solutions,
not technology
Emphasize projects that
improve customer travel
Projects
Empower end users with
insights, not data
Leverage continuous
improvement for apps
Use cloud to innovate
around legacy systems
Partner to acquire and
improve data
Focus on improving context
around data
Standardize formats for
consumption of insights
24
Questions
1. Does your enterprise have a comprehensive
cognitive program in place today?
2. When will your cognitive investment go beyond
the chat-bots and robots?
3. Are your cognitive investments more likely to
bring about a transformation in travel
personalization or operations?
4. How well is your company guiding the projects
that comprise its cognitive transformation
program?