Post on 12-Jan-2017
transcript
New Ways of Handling Old DataTom Vavra
AVP Software & Industry Insights and AnalysisIDC
2
Top Trends
© IDC Visit us at IDC.com and follow us on Twitter: @IDC
People Centric NetworksIntersection of People to People to DataCognitive Computing & Assistive Technology Work Context & flow
Organizational DynamicsSales, marketing, service not “working”Shifting workforce dynamics (for the 1st time
Millennials are the largest % of the workforce) Disruptive new connected business models
Applications are ChangingSocial: Inherent Ability to ConnectUnderlying Platform Services Distributed Information Access – Decision
Support
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 3
Unstructured Content: Value Waiting to be Delivered
Unstructured content – email, video, instant
messages, documents, and other formats –
accounts for
of all digital information
Unlocking value from this content should be
the goal of every organization, but very
few are actually getting all the value
they should be.
THIS CONTENT IS LOCKED IN A VARIETY LOCATIONS AND APPLICATIONS MADE UP OF SEPARATE REPOSITORIES THAT DON’T TALK TO EACH OTHER – E.G., EMC
DOCUMENTUM, SALESFORCE.COM, GOOGLE DRIVE, SHAREPOINT, ET AL.
90%
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 4
IDC’s Big Data and Analytics Predictions (1)
1. Through 2020, spending on cloud-based BDA technology will grow 4.5x faster than spending for on-premises solutions; open source technology will represent the core of this new architecture.
2. By 2020, 50% of all business analytics software will incorporate prescriptive analytics built on cognitive computing functionality.
3. Shortage of skilled staff will persist and extend from data scientists to architects and experts in data management; big data–related professional services will have a 23% CAGR by 2020.
4. By 2020, 90% of databases (relational and non-relational) will be based on memory-optimized technology.
5. By 2020, distributed micro analytics and data manipulation will be part of all big data and analytics deployments.
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 5
IDC’s Big Data and Analytics Predictions (2)
5. Through 2020, spending on self-service visual discovery and data preparation market will grow 2.5x faster than traditional IT-controlled tools for similar functionality.
6. By 2020, data monetization efforts will result in enterprises pursuing digital transformation initiatives increasing the marketplace’s consumption of their own data by 100-fold or more.
7. By 2020, the high-value data part of the digital universe that is worth analyzing to achieve actionable intelligence will double.
8. By 2020, 60% of information delivered to decision makers will be considered by them always actionable, doubling the current rate.
9. By 2020, organizations able to analyze all relevant data and deliver actionable information will achieve an extra $430 billion in productivity benefits over their less analytically oriented peers.
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 6
Why Do So Few Organizations Find Value in Their Information?
of knowledge workers regularly access 4 or more systems to get the information they need to do their jobs
of a typical knowledge worker’s day is spent looking
for and consolidation information spread across a
variety of systems
61% 36%
Nearly 15% access 11 or more systems
These workers find the information required to do their jobs only 56% of the time
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 7
Cognitive Software Attributes
• Performs deep natural language processing and analysis both for information ingestion and research as well as to provide human style communication (usually posed as questions and answers)
• Conducts learning in real time as data arrives • Has the ability to identify similar past experiences and use learning to
current situation • Predicts and recommends possible outcomes
• Score those outcomes with evidence for human analysis• Cycle back to the start so that the continuous learning is practiced, making
the system better over time
Cognitive software support human decision-making with more accuracy, confidence, speed, and agility based on broader and deeper bodies of evidence applied to a more comprehensive view of pertinent conditions without bias.
8
The Content Analytics, Discovery & Cognitive Systems Market Defined
Content Analytics• Text Analytics, Video Analytics• Categorizers and clustering engines• Speech Recognition, Language analyzers
Discovery• Enterprise search engines, information access
platforms, and applications for browsing and navigation
• Knowledge Base/Graph Generation• Rich media search
Cognitive Systems• Digital assistants• Automated advisors• Artificial intelligence, deep learning and machine
learning• Automated recommendation systems
© IDC Visit us at IDC.com and follow us on Twitter: @IDC
Clustering Engines
Speech Recognition
Video Analytics
Text Analytics
Kno
wle
dge
Bas
es &
G
raph
s
Ent
erpr
ise
Sea
rch
&
Info
rmat
ion
Acc
ess
Automated Advisors
Artificial Intelligence, Deep Learning and Machine
Learning
Clo
udO
n P
rem
ise
Commercial
Open Source
CADCS software analyzes, organizes, accesses, and provides advisory services based on a range of unstructured information and provides a platform for the development of analytic and cognitive applications.
9
Cognitive Solutions Ecosystem
Source: IDC
Behavioral Interactional
Performance
Long form
Geolocation
News
Personal data
Healthcare
Location
Sports & Entertainment
SocialCorporate
Logistics
Financial
Marketing
Sales
Procurement
Asset mgmt.
R&D
Logistics
HR
Anti money laundering
Retail pricing
Patient outcomes
Telco churn
IT performance mgt.
Retail
Travel
Media
Healthcare
Insurance
Investment
Commercial leasing
Advertising
Legal
Driverlesscars
Smart home devices
Self-flying drones
Robotic systems
Text analysis
Video analysis
Image analysis
Predictive analytics
NLP
APIs
ConnectorsData storesHypotheses generation
Machine learning
Speech Recognition
Dialogue Mgt.
Finance
Risk mgmt.
Weather
10
Cognitive Systems Use Cases Healthcare
• Diagnosis and Treatment Systems• Education and Training Systems• Pharmaceutical Research and Discovery
Retail• Expert Shopping Advisors & Product Recommendations• Automated Customer Service Agents• Automated Training Systems
Finance/Insurance• Automated Financial Advisors• Policy Advisors & Question and Answer Systems• M&A Investigation and Recommendations
Government• Police Investigation Systems• Program Advisors and Recommendation Systems
Manufacturing• Operational Improvement Systems• Asset Maintenance Systems
11
Source: IDC, 2015
2014–2019 Revenue ($M) with Growth (%)($M) (%)
2014 2015 2016 2017 2018 20190
5001,0001,5002,0002,5003,0003,5004,000
051015202530354045
827 1075 14191916
2644
3683
Cognitive Total growth (%)
Game Changer
Commercial cognitive software platforms have just begun to emerge on the market scene. This category of software used to build “smart” applications and expert advisors will grow rapidly over the next five years enabling a multi-billion dollar intelligent applications market.
2014–2019 Revenue ($M) with Growth (%)($M) (%)
Game Changer
Commercial cognitive software platforms have just begun to emerge on the market scene. This category of software used to build “smart” applications and expert advisors will grow rapidly over the next five years enabling a multi-billion dollar intelligent applications market.
Worldwide Cognitive Software Platform Forecast
Worldwide Cognitive Market by Industry, 2015
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 13
WW Cognitive Systems Spending (US$M) by Use Case, 2015
Adaptive Learning
Asset/Fleet Management
Automated Claims Processing
Automated Customer Service Agents
Automated Threat Intelligence and Prevention Systems
Defense, Terrorism, Investigation and Government Intelligence Systems
Diagnosis and Treatment Systems
Expert Shopping Advisors & Product Recommendations
Fraud Analysis and Investigation
Freight Management
Merchandising for Omni Channel Operations
Others
Pharmaceutical Research and Discovery
Program Advisors and Recommendation Systems
Public Safety and Emergency Response
Quality Management Investigation and Recommendation Systems
Regulatory Intelligence
Sales Process Recommendation and Automation
$- $ 100 $ 200 $ 300 $ 400 $ 500 $ 600 $ 700
Value (USD M)
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 14
European Cognitive Systems Spending (US$M) by Use Case, 2015
Adaptive Learning
Asset/Fleet Management
Automated Claims Processing
Automated Customer Service Agents
Automated Threat Intelligence and Prevention Systems
Defense, Terrorism, Investigation and Government Intelligence Systems
Diagnosis and Treatment Systems
Expert Shopping Advisors & Product Recommendations
Fraud Analysis and Investigation
Freight Management
Merchandising for Omni Channel Operations
Others
Pharmaceutical Research and Discovery
Program Advisors and Recommendation Systems
Public Safety and Emergency Response
Quality Management Investigation and Recommendation Systems
Regulatory Intelligence
Sales Process Recommendation and Automation
Supply and Logistics
$- $ 20 $ 40 $ 60 $ 80 $ 100 $ 120
2015
2
3
9
11
12
1
Source: IDC, 2016
Practices to Implement Cognitive Systems Initiatives
Source: IDC, 2016
1 - Setting Expectations
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 17
Be Realistic
Issues Business and IT both assume Cognitive will
replace humans Cognitive can only assist Outputs are never a “sure bet” Requires collaboration between IT and LOBs Relevant data is needed, not just more of it
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 18
Real World #1: Bankers vs. Robots
Source: IDC, 2016
2 – Leverage Cloud Services
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 20
Cloud as a Facilitator and Problem Solver
Issues Cognitive systems require vast data and
processing power On premise investment can be expensive and
time consuming Cloud services can do “heavy lifting” and
alleviate up front costs and time… … but, not all Cognitive solutions are Cloud-
ready or appropriate
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 21
Real World #2 :When Planes Love Clouds
We Cloud
Source: IDC, 2016
3 – Identify Repetitive Routine Actions
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 23
Choosing a Starting Point
Issues Identifying the right use case takes time and
thought Need to start by documenting current business
processes to identify resource-intensive tasks Narrow down list for cognitive applications Free up human resources to analyze exceptions
and outliers
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 24
Real World #3 : Letting Doctors be Doctors
Source: IDC, 2016
4 – Validate Outputs
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 26
Cognitive systems don’t fix bad inputs and untrained users
Issues The lack of quality inputs and expert trained
users will, necessarily, result in mistakes and bad outputs
Surprises are common in the early stages Constant validation is required to minimize
erroneous results Feedback on errors must be part of the regular
workflow
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 27
Real World #4 : Matching Clients with Hotels
Source: IDC, 2016
5 – Manage Data to Avoid Inaccurate Results
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 29
Data management is always central and key to any initiative
Issues Data can have many issues: inconsistency,
varied formats, ownership, (lack of) governance Need to map data: Where is it? Who owns it? Is
it connected/integrated? Third-part data is often required to complement
existing sources Value of data increases exponentially when
different types and sources are combined
© IDC Visit us at IDC.com and follow us on Twitter: @IDC 30
Real World #5 : IT Mfg + Internal IT + External provider
Thank you!
Tom Vavra
Tel: + 420 221 423 140
tvavra@idc.com
Associate Vice President
IDC CEMAMalé naměstí 13110 00 Praha 1Czech Republic
www.idc-cema.comwww.idc.com
CEMA Region