WHY ARTIFICIAL INTELLIGENCE - Accenture Insights · WHY ARTIFICIAL INTELLIGENCE FUTURE OF E GROWTH Dennis Kersten Kees van Mansom Accenture Innovation Deepdive 4th of July 2017. 2.
Post on 19-May-2020
7 Views
Preview:
Transcript
WHY ARTIFICIAL
INTELLIGENCE
FUTURE OF
GROWTHIS
T
HE
Dennis Kersten
Kees van Mansom
Accenture Innovation Deepdive
4th of July 2017
2
MEET ALEXA
Your assistant for Today!
3
4
TECHNOLOGY
WAVES
Mainframe
Client-Server and PCs
Web 1.0 eCommerce
Web 2.0, Cloud, Mobile
Big Data, Analytics, Visualization
IoT and Smart Machines
Artificial Intelligence
Quantum Computing
Quantum
ARTIFICIAL
INTELLIGENCE
JOURNEY
• Advances in machine learning, coupled with big data and cheap, ubiquitous cloud computing will unleash remarkable new potential for organizations across industries
• Initial research focused on AI, "the science aiming to create intelligent machines that are as capable as humans.”
• Recently, the focus has shifted from AI to Intelligence Augmentation (IA), intelligent systems that can support humans in their activities.
5
6
SENSE. INTERACT NATURALLY
HOW DOES
ARTIFICIAL
INTELLIGENCE
WORK?
COMPREHEND.WITHOUT ALL BUSINESS RULES DEFINED
ACT. EXTEND BEYOND HUMAN
LEARN. INSTEAD OF CODING
6
AI TECHNOLOGIES,
CAPABILITIES AND SOLUTIONS
Natural Language Processing
Machine translation
Emotion detection
Language detection
Content classification
…
Language
Computer vision
Object recognition
Face recognition
Object tracking
Optical Character Recognition
Handwriting recognition
Emotion detection
Gender/age detection
Scene recognition
…
Vision
Audio processing
Speech To Text/ Diarization
Speech recognition
Text to speech
Emotion detection
Language detection
Sound recognition
Gender/age detection
Speaker detection
…
Sound
Cap
ab
ilit
ies
Machine Learning, Deep Learning
So
luti
on
s
Hig
h P
erf
orm
an
ce B
usin
ess P
rocesses
Data
/Kn
ow
led
ge/E
xp
eri
en
ce
Unlock Value
in Dark Data
Intelligent
Process
Automation
Augmented
Intelligence
Open Source, Platforms (Google, Microsoft, Watson, Facebook, Amazon…), Specialist Products
Intelligent
Enterprise
Evolution
Strategy
Tech
no
log
ies
Virtual
Agents
Text
Analytics
Conversational
Intelligence
Software
Robotics
Video
Analytics
Unique
Identity
Caseload
Analytics
7
8
AI BUSINESS VALUE FORMULA
Reimagine
Business
Models and
Processes
Transform
Relationship
between
Human and
Machines
Unlock
Trapped Value
of Data
8
BENEFITS OF ARTIFICIAL
INTELLIGENCE
Increased productivity with
the potential to operate 24/7.
Less FTEs needed to
complete repetitive tasks
Approximately 6 weeks
required for a cost effective
implementation
40%
AI/Robotics can deliver
payback on investment
in 3 - 6 months when
implemented at scale
Higher staff satisfaction by eradicating
monotonous tasks allowing individuals
to focus on higher value work
43%43% increase in
FTEs able to focus
on customer
outcomes,
eliminating human
error
43%
Seasonal demand can be
managed by deploying virtual
resources at a fraction of the
cost of an FTE
Consistent quality
guaranteed as human error is
eliminated
Processing costs
reduced by up to
80%80%
Provision of greater visibility and
auditability of transactions, leading
to better control over end to end
process. Average handling
times reduced by
40%, with a 24/7
resilient operation
9
COLLETTE:
MORTGAGE
ADVISOR OF THE
FUTURE
10
Natural Language
Processing
Machine translation
Emotion detection
Language detection
Content classification
…
Language
Audio processing
Speech To Text/ Diarization
Speech recognition
Text to speech
Emotion detection
Language detection
Sound recognition
Gender/age detection
Speaker detection
…
Sound
Cap
ab
ilit
ies
Machine Learning, Deep Learning
So
luti
on
s
Augmented
Intelligence
Platforms (IBM Watson), Nuance Nina
Tech
no
log
ies
Virtual
Agents
Conversational
Intelligence
Caseload
Analytics
1111
VIDEO
ANALYTICS:
MAKING VIDEO
SURVEILLANCE
SMARTER 12
Computer vision
Object recognition
Face recognition
Object tracking
Optical Character
Recognition
Handwriting recognition
Emotion detection
Gender/age detection
Scene recognition
…
Vision
Audio processing
Sound recognition
Sound
Cap
ab
ilit
ies
Machine Learning, Deep Learning
So
luti
on
s
Unlock Value
in Dark DataAugmented
Intelligence
Open Source, Platforms (Microsoft,), VCA
Tech
no
log
ies
Video
Analytics
1313
DEEP LEARNING:
AUTO INSURANCE
CLAIMS
PROCESSING
14
Computer vision
Object recognition
Vision
Cap
ab
ilit
ies
Machine Learning, Deep Learning
So
luti
on
s
Intelligent
Process
Automation
Augmented
Intelligence
Open Source
Tech
no
log
ies
Caseload
Analytics
Copyright © 2016 Accenture All rights reserved. 15
Case Study: Auto Insurance Claims Processing Automate Classification of Car Damage Level
Given only an image, classify a car as:
• Undamaged
• Damaged
• Totaled
Problem Details
•An Insurance Company wanted to automate
claims processing using advanced machine
learning technology, namely Deep Learning.
•When customers sent a picture of their
damaged car, the Company would like to have
the ability of automatically detect the level of
damage and use it to, for example, order spare
parts and possibly detect fraud, if any.
•Accenture developed a Convolutional Neural
Network algorithm (which belongs to the family
of Deep Learning techniques) using a data set
of toy images.
Copyright © 2016 Accenture All rights reserved. 16
Key Technology: Deep Learning
Computer science: The learning machines Nature News & Comment, 2014
Deep Learning: different layers of abstraction inspired by human cognition
Image Source Maurice Peemen
Copyright © 2016 Accenture All rights reserved. 17
Training Data – 500 Images per Damage Level Set
Undamaged cars Damaged cars Totaled cars
Copyright © 2016 Accenture All rights reserved. 18
Damage Classification Results
90% accuracy achieved• By automatically detecting level of damage,
an Insurance Company saves on sending a
human to assess the damage
• Apply the same technique for more use
cases and other lines of businesses like
Home Insurance with enhanced complexity
and accuracy
• For Auto Insurance, spare parts could be
ordered automatically
• For Home Insurance, evaluate building
resistance, check if customers are telling
the truth about additions to houses, identify
multiple damages
• Similarities/differences in damage patterns
could be used to detect fraud
Value Delivered
B2C
CUSTOMER
ADVISOR
19
Natural Language
Processing
Machine translation
Emotion detection
Language detection
Content classification
…
Language
Audio processing
Speech To Text/ Diarization
Speech recognition
Text to speech
Emotion detection
Language detection
Sound recognition
Gender/age detection
Speaker detection
…
Sound
Cap
ab
ilit
ies
Machine Learning, Deep Learning
So
luti
on
s
Augmented
Intelligence
Platforms (IBM Watson), Nuance Nina
Tech
no
log
ies
Virtual
Agents
Conversational
Intelligence
Caseload
Analytics
20
THE INTELLIGENT AUTOMATION VALUE SPECTRUM
Complexity
Cog
nitiv
ity
+
+
Robotic Process Automation (RPA)
Artificial Intelligence / Cognitive Reasoning
Dynamic Process Orchestration
Workforce Management
21
THE INTELLIGENT AUTOMATION VALUE SPECTRUM
Complexity
+
+
Robotic Process Automation (RPA)
Artificial Intelligence / Cognitive Reasoning
Dynamic Process Orchestration
(Virtual) Interaction
Workforce ManagementCog
nitiv
ity
22
• Data Complexity: degree to which complex unstructured changing data needs to be taken into account
• Task Complexity: degree to which individuals need to apply their judgment and interpret a variety of information
TURNING ARTIFICIAL INTELLIGENCE
INTO BUSINESS VALUE
Task Complexity
EFFECTIVENESS MODELL
Support seamless integration and
collaboration
• Wide range of interconnected work activities
• Highly reliant on coordination and communication
• Virtual Agents
INNOVATION MODEL
Enable creativity and ideation
• Original, innovative work
• Highly reliant on deep expertise
• Support for biomedical research, fashion design, music
written
EFFICIENCY MODEL
Provide consistent, low cost performance
• Routine work
• Reliant on well-defined and well-understood criteria
• Automated credit decisions
EXPERT MODEL
Leverage specialized expertise
• Judgement-oriented work
• Reliant on expertise and experience
• Expert system medical diagnoses or financial research
Unstructured, Volatile,
High-Volume
Structured, stable, low
volume
Routine, Predicable, Rules,
based
Ad Hoc, Unpredictable,
Judgement-based
Data Complexity
AUTOMATE
AUGMENTED
23
TRANSFORMING THE IT ORGANIZATION
AND ARCHITECTURE
Live Agents
(escalation/
Level 2+)
Back end systems
(ERP, CRM, SaaS,
mainframes, custom…)
Devices
Phone
Tablet
Mobile
SMS
PC
Channels
Calls
/fax
App
Chat
Web
Users Auto Agents
(Auto response/
routing/Level 1)
Robotics
(process
execution)
24
THE RIGHT APPROACH
Data
System interfaces
Process steps
Business Rules
Goals
•Traditional Business
Processes are modeled from
a process perspective;
outside – in
• Intelligent Automation
solutions are designed with
the end goal in mind; inside –
out
25
26
COMPETENCES TO SUCCEED
Current
stage
Desired
stage
0 1 2 3 4 Knowledge
Skills
Abilities
26
27
APPROACH & TIME LINES
DiscoveryProof of Value Design &
DevelopmentProof of Value Implementation
Pilot /
Production
1 week 1 month 1 – 2 months
1 week 1 month 2 – 4 months
2 weeks 1 month 2 – 4 months
2 weeks 1 month 2 – 4 months
Straight-through
processing
Straight-through
processing
Amplified Hybrid
Workforce
Optimized
customer journey
THE BEST WAY TO
PREDICT THE FUTURE
IS TO INVENT IT
28
DENNIS KERSTENAI lead for Netherlands
Accenture
CONTACT US
www.accenture.com/futureofAI
KEES VAN MANSOMNL lead Automation
Accenture
29
ENTERPRISE
ARCHITECTURE
CULTURE
DATA
WHAT’S NEXT
FOR AI?INCREMENTAL VS
TRANSFORMATIONAL
ENTERPRISE
CHALLENGES
30
top related