Purpose: The slides provide an overview on the AI/Cognitive Computing trend
Content: AI/Cognitive Computing market trends, drivers, spending trends, industry business cases, and adoption challenges. Also included are links to additional resources.
How To Use This Report: This report is best read/studied and used as a learning document. You may want to view the slides in slideshow mode so you can easily follow the links
Available on Slideshare: This presentation (and other Trend Reports for 2017) will be available publically on Slideshare at http://www.slideshare.net/horizonwatching
Please Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Dev Dives: Streamline document processing with UiPath Studio Web
AI / Cognitive Computing Trend Report, 2017
1. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Welcome to the new era of Cognitive Computing
Andrew Onda, Manager – Cognitive Solutions, Market Development & Insights
Bill Chamberlin, Distinguished Market Intelligence Professional, MD&I HorizonWatch
February 15, 2017
AI / Cognitive Computing
Trend Report, 2017
2. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
About This Trend Report
15Feb2017
Purpose: The slides provide an overview on the AI/Cognitive Computing trend
Content: AI/Cognitive Computing market trends, drivers, spending trends,
industry business cases, and adoption challenges. Also included are links to
additional resources.
How To Use This Report: This report is best read/studied and used as a
learning document. You may want to view the slides in slideshow mode so you
can easily follow the links
Available on Slideshare: This presentation (and other Trend Reports for
2017) will be available publically on Slideshare at
http://www.slideshare.net/horizonwatching
Please Note: This report is based on internal IBM analysis and is not meant to
be a statement of direction by IBM nor is IBM committing to any particular
technology or solution.
AI / Cognitive Computing Trend Report, 2017 (External Version)2
3. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Welcome to the Cognitive Era and the rise of
Cognitive Business
3
We have entered a new era of computing
called cognitive computing. It signifies a
fundamental shift in how machines interact
with us, other machines and the environment.
It will provide much value, but it will be highly
disruptive.
The Cognitive Era will result in entirely new
model of computing that includes a range of
technology innovations in analytics, natural
language processing and machine learning
“Cognitive computing is quickly
emerging as a transformative technology
that enables organizations to gain
business advantage. This shift to
systems that can reason and learn is
especially germane to the bottom line;
the cognitive era is here in large part
because it makes practical business
sense.” IBM
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Key Insights
IBM: How Watson Works (video)
4. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Cognitive computing combines AI technologies,
relevant content and deep subject matter expertise
4
Capability Required Advanced Ultimate
Understand
domains deeply
• Has task-specific
knowledge
• Also has contextual
knowledge relevant to
asks
• Also has common sense
& general knowledge
Reason towards
specific goals
• Uses specific
approach
• “Deductive or
Inductive”
• Also explains
• “Abductive”
• Also combines
approaches
• “Adaptive”
Learn continuously
from experience
• Uses explicit training
• “Supervised learning”
• Also uses implicit
training.
• “Unsupervised
learning”
• Also uses feedback from
the environment.
• “Reinforcement learning”
Interact naturally
with cognitive
agents
• Transactional request
response
• Session-oriented
contextual, affective
interactions
• Life-long conversational
interactions
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Key Insights
5. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Cognitive Computing – Trend Overview
5
Drivers
Advances in enabling AI technology areas
Increasingly large complex datasets that can be the source of major
insights and hold the answers to critical questions
Third computing platform – Cloud, IoT, Mobile, Big Data, Analytics,
Social
Challenges
Much more work on enabling technologies required
Documented use cases and ROI impact not yet fully understood
Education future of AI’s impact on jobs and society
Window of opportunity: Partnerships / Ecosystems / Developers
Implications
First movers will have an advantage.
This is a technology that learns. Accuracy will increase over time.
Cognitive services will be delivered via the cloud to mobile device
Expect the mass market to develop around consumers at a fast pace.
Business processes and whole industries will be transformed.
Changes will be required in employee workforce, corporate culture,
and partner ecosystems.
New IT architectures, systems design, data management/analytics
and workload optimized systems.
Job displacement will happen, but new jobs will be created. Scenario
planning needed to understand long term implications.
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Market Trends
“The cognitive era will create new jobs, such
as robot monitoring professionals, data
scientists, automation specialists, and content
curators: Forrester forecasts 8.9 million new
jobs in the US by 2025.” Forrester
“Widespread adoption of cognitive systems
and artificial intelligence (AI) across a
broad range of industries will drive
worldwide revenues from nearly $8.0
billion in 2016 to more than $47 billion in
2020” IDC
World of Watson: Ginni Rometty keynote
6. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Top trends to watch in 2017 include vertical industry
use cases, chatbots and the impact of consumer AI
6
Verticals: Healthcare, Life Sciences, and FSS AI start-
ups are using machine learning and both structured and
unstructured data. customer facing departments are
leading or evaluating the investments and adoption in AI
systems. Enterprise buyers are partnering with AI
software providers as they start their AI journey
Conversational services explode: In 2017, expect
increased activity in conversational services (e.g.
chatbots, virtual agents). Vendors with best UX will
succeed. Apps & services must be easy to use and the
UX needs to be immersive, ambient, conversational and
accurate. It must also learn our preferences from past
conversations.
Consumer AI tech drives market. AI will be used by
consumers everywhere: IoT devices, home devices,
robots, vehicles, drones, digital assistants and
wearables. As usage grows, it will influence how B2B
markets develop with AI solutions migrating into the
enterprise, starting with the automotive, electronics, and
banking industries
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Market Trends
Forrester: Predictions 2017: Artificial Intelligence
Will Drive The Insights Revolution
“Near-term opportunities for cognitive systems
are in industries such as banking, securities and
investments, and manufacturing” Jessica
Goepfert, IDC
IBM: The Future of Cognitive Services
7. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Vertical industry use cases include diagnosis,
conversational services, and recommendation systems
7
Industries Near-Term Use Case Examples
Banking / FM / Insurance
Sales, Marketing, Service, Regulatory, Claims,
Risk
• Program Advisors and Recommendation
• Fraud Analysis and Investigation
• Regulatory Intelligence
• Automated Claims Processing
• Automated Threat Intelligence and Prevention Systems
Healthcare & Life Sciences
Services and decision support
• Diagnosis and Treatment
• Pharma R&D
• Claim Management
• Personalized Health Services
Consumer (Retail, CPG)
Sales, Marketing, Merchandising, Supply chain
logistics
• Automated Customer Service Agents
• Expert Shopping Advisors and Product Recommendations
• Omni-channel merchandising
Telco, M&E
Sales, Marketing, IT
• Virtual Agents
• Digital Advertising
• Automated Threat Intelligence
Government, Education
Core operations, Threat intelligence
• Threat Intelligence and Prevention
• Defense, Terrorism, Investigation
• Program Advisors and Recommendation Systems
• Public Safety and Emergency Response
• Adaptive Learning
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Market Trends
8. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Technologies to watch include neural networks,
machine learning, NLP, and pattern recognition
Neural Networks: Neural networks are providing rapid
advances in speech and image recognition and machine
translation. some systems perform as well as humans
Advancements in machine learning feed off access to IoT-
generated data. As the IoT generates more real-time data,
machine learning algorithms will use that data for planning,
forecasting and making critical business decisions.
Natural language processing (NLP) sweet spot is
customer service. Enterprises will use NLP-based solutions,
including chatbots, to handle customer service inquiries
without human interaction, improving overall efficiencies and
costs.
Rich media & IoT analytics fuel pattern recognition
advancements. Combining historical and traditional data with
real time rich media & IoT unstructured data will reveal new
patterns in data that can inform future decision processes.
Knowledge representation and reasoning efforts focus on
supporting decision making. Look for efforts to build large-
scale knowledge databases that cognitive systems can use to
drive unbiased support for subject-matter experts.
8 15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Market Trends
“Cognitive advances — combined with
robotics, IoT, artificial intelligence,
augmented reality/virtual reality, machine
intelligence, and massive data sets —
increase the ability of systems to mimic
and enhance human intelligence in real
time.” IDC
“By 2018, deep learning (deep neural
networks) will be a standard component
in 80% of data scientists’ tool boxes..”
Gartner
IBM Research: Cognitive Computing
9. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
The AI startup and vendor ecosystem will continue to
grow and expand in in 2017
9
VCs and Startups: The market landscape has 2,000+ start-
ups. Venture Capital funding was nearly $4.5B in 2016;
investments in core AI technology, chatbots, and Advanced
Analytics have grown over the last two years. The majority of
AI start-ups are using machine learning and both structured
and unstructured data
Large vendors continue to acquire talent and
capabilities. Continuing a trend we have seen the last three
years, expect all large vendors to look to acquire AI startups
and capabilities in order to build economies of scale and
fend off the competition. AI acquisitions have been strong
over the past 3 years, with a focus on machine learning
capabilities; Virtual Agents / Chatbots acquisitions were also
strong last year
Vendors continue to court developers to their platforms.
To accelerate innovation and monetize assets, platform
players will increasingly target development communities.
Patents: AI patent applications peaked in 2014 with a
significant drop in 2016. IBM has been the most active in
regard to patenting core AI technology. Look for increased
patent activity for niche and industry vertical applications.
“The startups have raised $3.8B in aggregate
funding across 263 deals since 2012, and
include startups at different investment
stages of development, from seed/angel
companies to well-funded unicorns.” CBInsights:
AI 100: The Artificial Intelligence Startups Redefining
Industries
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Market Trends
VentureBeat: 5 deep learning startups
to follow in 2017
“From stopping cyberattacks to operating
autonomous vehicles to visually searching
through a wine database, over 550 startups
using AI as a core part of their products raised
$5B in funding in 2016.” CB Insights
10. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
A recent survey of early adopters indicate there are
real business benefits of cognitive computing
10
65% say adopting cognitive is very important to their
organization’s strategy and success
58% say cognitive computing is essential to digital
transformation
50% say they already gain major competitive
advantage from their cognitive initiatives
53% say cognitive computing will unlock the hidden
value of their organization’s dark data
49% say it helps improve customer engagement and
customer service
49% say it has improved productivity & efficiency
42% say it is enabling business growth through an
expanded ecosystem
“Organizations already gain major
competitive advantage from their use
of cognitive computing. They achieve
a range of business outcomes–from
customer engagement to productivity
& efficiency and business growth.”
IBM: The cognitive advantage report
IBM: The cognitive advantage report
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
IBM: The cognitive advantage report
Market Trends
11. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Top Challenges to Adoption
11
1. Implementation effort / cost. In some cases, the effort and
cost to implement AI technologies can be high. Enterprises
want to understand the business value and ROI. Limited
case studies exist for cognitive computing, creating potential
hurdles to large-scale investment.
2. Data management and security / privacy Concerns. .
Cognitive systems will rely heavily on data derived from
multiple systems and sources. As such, it is important to
ensure data is consolidated and sourced in a consistent,
secure manner. Safeguards are needed against data
overreach and abuse of privacy.
3. Insufficient Skills. AI technologies and cognitive computing
solutions are emerging at a fast pace. Like any emerging
tech area, there is limited skills in the marketplace.
Lack of industry definitions and standards. Similar to
other emerging areas, the lack of industry standards and
definitions around artificial intelligence and cognitive solutions
may create hesitation in adoption.
Limitations of current technology. While there are
important reasons to move forward with cognitive technology,
firms must consider the constraints of existing platforms
relative to their ultimate goals.
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Adoption Challenges
“93% of automation technologists feel
unprepared or only partially prepared to
tackle the challenges associated with smart
machine technologies.” Forrester
O’Reilly: 3 challenges for artificial
intelligence in medicine
The Verge: These are three of the
biggest problems facing today's AI
12. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Selected IBM websites
12
IBM Website IBM Watson
Watson APIs
Products
Blog: Built with Watson
Watson Internet of Things
Watson Health
IBM DeveloperWorks Cognitive Computing
IBM Research: Cognitive Computing / DeepQA
White paper: Computing, cognition and the
future of knowing
IBM DeveloperWorks Cognitive Computing
IBM Website IBM Watson
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Featured Links
13. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Example Cognitive Computing Use Cases
13
1. MD Anderson Cancer Center (Healthcare) Solution that aggregates large volumes of unstructured patient
data from a variety of sources, enabling clinicians and researchers to run analytics in near-real time to identify
patterns and gain important insights.
2. The Chinese University of Hong Kong (Healthcare) Cloud-based solution that enables research teams to
accelerate their exploration of large, complex cancer data sets and identification of cancer trends.
3. Baylor College of Medicine (Education) Cognitive technology is used to accelerate medical research by
analyzing over 300,000 articles automatically, looking for key words and phrases that indicate correlations
among studies.
4. Guiding Eyes (Education) Solution that advances the art and science of raising guide dogs by analyzing
structured and unstructured data to discover genetic, health, temperament and environmental factors that
correlate with success.
5. Alpha Modus (Financial Markets) Cloud-based solution that can analyze large volumes of unstructured and
natural language data from financial services industry sources, including stock exchanges, social media, and
other market indicators, making it easy to explore new ideas and turn them into practical investment
applications.
6. WayBlazer (Travel & Transportation) Travel application and APIs that use analyze unstructured social media
content from thousands of sites and deliver personalized recommendations to travelers who issue natural-
language questions.
7. Honest Cafe (Retail) Analytics are used to explore vending machine transactions, payments and weather
data, find correlations, and uncover patterns in customer behavior.
8. Pon Holdings (Automotive, Industrial Products) Solution that extracts and analyzes unstructured data from
millions of web pages, using the insights to generate leads for sales and marketing departments and also to
improve pricing.
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Featured Links
14. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Selected public-facing analyst websites and resources
14
“The main driver for global cognitive computing market is increasing trend of large and complex data-set.
This complex data is available in various form such images, video, and natural language. These data-sets
cannot be processed with help of existing big data analytics systems, hence business organizations are
turning their heads towards the cognitive computing.” MarketsandMarkets
“Smart machines are no longer science-
fiction; they are here, being used by
businesses right now and only growing in
capabilities. The disruption they will
cause over the next decade will leave no
industry untouched.” Gartner
Forrester: Artificial Intelligence (blogs) and
Cognitive Computing (search)
Frost & Sullivan: Artificial Intelligence
(search)
Gartner: Smart Machines (search) and
Machine Learning
IDC: Intelligent Systems (search) and
Machine Learning and Cognitive
Computing
MarketsandMarkets: Cognitive
Computing
Gartner Press Release: Smart Machines Will Enter
Mainstream Adoption By 2021
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Featured Links
15. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Selected media websites
15
“Artificial intelligence suffered a long winter,
but a new name -- cognitive computing -- and
a flood of data, innovation, and compute
power now has thousands smart applications
flourishing.” Infoworld: It's springtime at last
for cognitive computing
“The opportunities for cognitive computing
are compelling. Organizations that move
forward with this innovative capability will
have the opportunity to push new boundaries
for growth and gain advantage over their
competition.” Forbes: Four Steps To Get
Started With Cognitive Computing
CIO.com Artificial Intelligence (search) / IBM Watson (search)
ComputerWorld Artificial Intelligence (search) / IBM Watson
(search)
DataVersity Cognitive Computing
eWeek Cognitive Computing (search) / Artificial Intelligence
(search)
Forbes Cognitive Computing (search) / Artificial Intelligence (search)
GIGAOM Artificial Intelligence (search
HBR.org Cognitive Computing (search) / Artificial Intelligence
(search)
InformationWeek Cognitive Computing (search) / Artificial
Intelligence (search)
InfoWorld Cognitive Computing (search) / Artificial Intelligence
(search):
Wired Artificial Intelligence (search) / IBM Watson (search)
Wired Artificial Intelligence is More
Artificial than Intelligent
HBR: How Artificial Intelligence Will
Redefine Management
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Featured Links
16. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Selected IBM social media resources
16
Twitter: @ibmwatson Hashtags #ibmwatson
Linkedin: IBM Watson Group
Facebook: IBM Watson
YouTube: IBMWatson
Instagram: Watson
Slideshare.net search for IBM Watson
Blogs: Watson Blog
Twitter: @ibmwatson
Blog: Watson Blog
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Featured Links
17. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
Selected links to additional articles and resources
CIO: 2017 will be big year for AI thanks to tech giants
Cognitive Computing Consortium: Where Are We With Cognitive Computing Today? Part 1
Cognitive Computing Consortium: Where Are We With Cognitive Computing Today? Part 2.
ComputerWorld: Artificial Intelligence: Transparency isn't just a trend
Forbes: Crawl With Analytics Before Running With Artificial Intelligence
Forbes: 10 Powerful Examples Of Artificial Intelligence In Use Today
IBM: Planning for 2017? Top 3 ways you can add business value with cognitive solutions
Information Age: How artificial intelligence will support business growth in 2017
O'Reilly: 7 AI trends to watch in 2017
Technology Review: 5 Big Predictions for Artificial Intelligence in 2017
Technology Review: Survey: Machine Learning Trends, Challenges, and Opportunities
VentureBeat: 3 big trends for AI in 2017
17 15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)
Featured Links
18. Note: This report is based on internal IBM analysis and is not meant to be a statement of direction by IBM nor is IBM committing to any particular technology or solution.
More Insights on Technology Trends are Available
18
Other slide decks in this 2017 Trend Report series have been posted to Slideshare
You are also invited to check out the following IBM websites and resources
– IBM Academy of Technology
– IBM Institute for Business Value
– IBM Research and Research News and 5 in 5
– IBM’s THINK blog
– IBM Think Academy on YouTube
15Feb2017 AI / Cognitive Computing Trend Report, 2017 (External Version)