Building a Cognitive Business
M A Y 2 0 1 7
We are at the beginning of the most
transformative revolution ever
Industry 1.0
Steam
Industry 2.0
Electricity
Industry 3.0
Information Age
Industry 4.0
Connected Intelligence
“I am telling you, the
world’s first trillionaires
are going to come from
somebody who masters
AI and all of its
derivatives and applies it
in ways we never thought
of”
- M A R K C U B A N
“We will move from
mobile first to an AI first
world.”
S U N D A R P I C H A I ,
G O O G L E C E O
I N D U S T R Y 1 . 0
Steam
I N D U S T R Y 3 . 0
Information Age
I N D U S T R Y 4 . 0
Connected Intelligence
150 Years
75 Years
40 Years
?? Years
I N D U S T R Y 2 . 0
Electricity
85%of a businesses’
customer relationships
will be managed by
virtual assistants by
2020
Source: Gartner
49%of all activities people
are paid to do today
have the potential to be
automated by adapting
currently demonstrated
technology.
Source: McKinsey
By 2020, the
average person
will have more
conversations with
bots than with
their spouse
Source: Gartner
$2 TrillionExpected market value to be created over the next decade
~60%Percentage that will come from incumbents
Source: Yuri Milner, Fortune Brainstorm Tech 2016
Source: Yuri Milner, Fortune Brainstorm Tech 2016
On average, an S&P company is being replaced
Every Two Weeks
Source Richard Foster, Yale School of Management
Don’t bring a sword to a gun fight
Don’t bring a 20th century
business model to an AI-first
world
C O P Y R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 9
1) Assess what is possible
2) Define your data needs
3) Define the services
4) Align to technology platforms
5) Execute real-world solutions
6) Incorporate experience design into
everything
7) Iterate, iterate, iterate…
Seven Steps of Cognitive
Transformation
C O P Y R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 10
CONVERSATIONAL
TECHNOLOGY
INSIGHT
GENERATION
BUSINESS
ACCELERATION
Step 1: Assess What Is Possible
The Three Pillars of a Cognitive Business
CHATBOT ASSISTANT EXPERT SAVANT
TODAY 2017-2018 2018-2020 2030+
Your customers
will EXPECT expert
digital assistants
within the next
five years
CONVERSATIONAL TECHNOLOGY
C O P Y R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 12
INSIGHT GENERATION
GROWTH
STRATEGY
JOURNEY
DESIGN
CREATIVE
BUYING &
OPTIMISATION
ATTRIBUTION &
MEASUREMENT
Identify the unknown Move the needle KPI-wise
DATA
C O P Y R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 13
Pretty much anything that a normal
person can do in <1 sec, we can now
automate with AI.”
- Andrew Ng
BUSINESS ACCELERATION
C O P Y R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 14
7.5%
3.0%
AI only Human only
Errors in diagnosing cancer in
lymph node cells
0.5%
AI + Human
BUSINESS ACCELERATION
C O P Y R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 15
BUSINESS ACCELERATIONINSIGHT GENERATIONCONVERSATIONAL TECHNOLOGY
• Call Center
• Optimized call routing
• Self service
• Digital Agent
• Top of funnel increases
• Increased conversion
• Claims Data Collection
• Detailed follow-up
• Email drafting
• Internal Help Desk
• Self service
• Increased employee
efficiency
• Agent Support
• Increased employee
efficiency
• Improved customer
satisfaction
• Policy Pricing
• Optimize against expected
outcomes
• Leverage third party data
• Hold vs Sell Optimization
• Leverage additional
information for hold/sell
• Advertising Expenses
• Optimized targeting
• Increased yield models
• Customer Insights
• Increase NPS score
• Proactively identify churn
• Product Development
• Identify new behaviour
patterns
• Claims Processing
• Automate basic tasks
• Reduce human error
• Fraud Detection
• Improve fraud detection
• Increased staff efficiency
• Client Onboarding
• Automate basic tasks
• Manage intake of documents
• Documentation processing
• Automate basic tasks
• Email drafting
• AR / AP Automation
• Automate report creation
• Automate collection activities
16
© Copyright Publicis.Sapient | Confidential
WHAT YOU KNOW
1st Party Data
WHAT THEY SAY
Social Behavior
WHAT THEY DO
Consumer Behavior
THEIR OPINIONS
Survey Data
Step 2: Define Your Data Needs
WHY
C O P Y R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 17
Step 3: Define the Services
Text Generation
Query Parser
Results
Post-Processor
Indexing
Structured Data
Access
Ontology Creation
Ontology Evolution
Facets/Filters
Relevance-based
ResultsRanking
UserEntitlements
Check
Tagging/Metadata
Extraction
Content Storage
Concepts/Entities
Extraction
Relations/Fac
t Extraction
Heuristic
Inference
Deductive
SERVICES API (KNOWLEDGE ENGINE ABSTRACTION LAYER)
Concepts
Disambiguation
Free-Text Search Contextual Search Graph Search
Reasoning /
Explanation
Query Formulation Query Expansion NLP / NLU
Chat Agents Voice & Tone Mobile Framework ASR / TTS
USER EXPERIENCE FRAMEWORK
Reporting / Portal Email
Text Interfaces Audio Capture
TEXT & AUDIO INTERFACES
Audio Output
REST / SOAP APIs
Managed File
Transfer
Security
Logging
Caching
Monitoring
UserEntitlements
REASONING
ENGINES
INTEGRATION &
CORE SERVICESKNOWLEDGE BASE MGMT DATA & CONTENT INGESTION DATA & CONTENT DELIVERY
Topic Modelling,
Query Exapansion,
NLP/NLU, Ontology
Indexes, Structured
Data Triples
Knowledge Base,
Reasoning
Natural Language
Translation
Deep Learning
Engines
Constrained
Conditional Models
Ontology Neural Networks
KNOWLEDGE SERVICES MACHINE LEARNING
C O P Y R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 18
Step 4: Align to Technology Platforms
Text Generation
Query Parser
Results
Post-Processor
Indexing
Structured Data
Access
Ontology Creation
Ontology Evolution
Facets/Filters
Relevance-based
ResultsRanking
UserEntitlements
Check
Tagging/Metadata
Extraction
Content Storage
Concepts/Entities
Extraction
Relations/Fac
t Extraction
Heuristic
Inference
Deductive
SERVICES API (KNOWLEDGE ENGINE ABSTRACTION LAYER)
Concepts
Disambiguation
Free-Text Search Contextual Search Graph Search
Reasoning /
Explanation
Query Formulation Query Expansion NLP / NLU
Chat Agents Voice & Tone Mobile Framework ASR / TTS
USER EXPERIENCE FRAMEWORK
Reporting / Portal Email
Text Interfaces Audio Capture
TEXT & AUDIO INTERFACES
Audio Output
REST / SOAP APIs
Managed File
Transfer
Security
Logging
Caching
Monitoring
UserEntitlements
REASONING
ENGINES
INTEGRATION &
CORE SERVICESKNOWLEDGE BASE MGMT DATA & CONTENT INGESTION DATA & CONTENT DELIVERY
Topic Modelling,
Query Exapansion,
NLP/NLU, Ontology
Indexes, Structured
Data Triples
Knowledge Base,
Reasoning
Natural Language
Translation
Deep Learning
Engines
Constrained
Conditional Models
Ontology Neural Networks
KNOWLEDGE SERVICES MACHINE LEARNING
Bespoke &
Point Solutions
Semantic
AI
Natural Language
& Vision Platforms
Align
with
Internal
Arch
Deep
Learnin
g
ML EngineKnowledge Engines (ML, Semantic, NLU)
C O P Y R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 19
KNOWLEDGE
LEVERAGE
SCALABILITY EXPERIENC
E DESIGN
SECURITYLEARNING
Step 5: Execute Real-World Solutions
BUILD AS AN INTEGRATED PLATFORM
NOT AS POINT SOLUTIONS
C O P Y R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 20
Step 6: Incorporate Experience Design
Happy Path
C O P Y R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 21
Step 6: Incorporate Experience Design
Happy Path
Sad Facts
Changed Mind
Make No Sense
Unexpected Question
Small Talk
C O P Y R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 22
Step 7: Iterate
DATA NEEDS
HUMAN BEHAVIOUR
TECHNOLOGY SHIFTS
LEARN FROM FAILURE
FIND NEW OPPORTUNITIES
Building a Cognitive Business – Josh Sutton, SapientRazorfish Data & Artificial Intelligence Practice Lead

Building a Cognitive Business – Josh Sutton, SapientRazorfish Data & Artificial Intelligence Practice Lead

  • 1.
    Building a CognitiveBusiness M A Y 2 0 1 7
  • 2.
    We are atthe beginning of the most transformative revolution ever Industry 1.0 Steam Industry 2.0 Electricity Industry 3.0 Information Age Industry 4.0 Connected Intelligence
  • 3.
    “I am tellingyou, the world’s first trillionaires are going to come from somebody who masters AI and all of its derivatives and applies it in ways we never thought of” - M A R K C U B A N “We will move from mobile first to an AI first world.” S U N D A R P I C H A I , G O O G L E C E O I N D U S T R Y 1 . 0 Steam I N D U S T R Y 3 . 0 Information Age I N D U S T R Y 4 . 0 Connected Intelligence 150 Years 75 Years 40 Years ?? Years I N D U S T R Y 2 . 0 Electricity
  • 4.
    85%of a businesses’ customerrelationships will be managed by virtual assistants by 2020 Source: Gartner 49%of all activities people are paid to do today have the potential to be automated by adapting currently demonstrated technology. Source: McKinsey By 2020, the average person will have more conversations with bots than with their spouse Source: Gartner
  • 5.
    $2 TrillionExpected marketvalue to be created over the next decade ~60%Percentage that will come from incumbents Source: Yuri Milner, Fortune Brainstorm Tech 2016 Source: Yuri Milner, Fortune Brainstorm Tech 2016
  • 6.
    On average, anS&P company is being replaced Every Two Weeks Source Richard Foster, Yale School of Management
  • 8.
    Don’t bring asword to a gun fight Don’t bring a 20th century business model to an AI-first world
  • 9.
    C O PY R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 9 1) Assess what is possible 2) Define your data needs 3) Define the services 4) Align to technology platforms 5) Execute real-world solutions 6) Incorporate experience design into everything 7) Iterate, iterate, iterate… Seven Steps of Cognitive Transformation
  • 10.
    C O PY R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 10 CONVERSATIONAL TECHNOLOGY INSIGHT GENERATION BUSINESS ACCELERATION Step 1: Assess What Is Possible The Three Pillars of a Cognitive Business
  • 11.
    CHATBOT ASSISTANT EXPERTSAVANT TODAY 2017-2018 2018-2020 2030+ Your customers will EXPECT expert digital assistants within the next five years CONVERSATIONAL TECHNOLOGY
  • 12.
    C O PY R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 12 INSIGHT GENERATION GROWTH STRATEGY JOURNEY DESIGN CREATIVE BUYING & OPTIMISATION ATTRIBUTION & MEASUREMENT Identify the unknown Move the needle KPI-wise DATA
  • 13.
    C O PY R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 13 Pretty much anything that a normal person can do in <1 sec, we can now automate with AI.” - Andrew Ng BUSINESS ACCELERATION
  • 14.
    C O PY R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 14 7.5% 3.0% AI only Human only Errors in diagnosing cancer in lymph node cells 0.5% AI + Human BUSINESS ACCELERATION
  • 15.
    C O PY R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 15 BUSINESS ACCELERATIONINSIGHT GENERATIONCONVERSATIONAL TECHNOLOGY • Call Center • Optimized call routing • Self service • Digital Agent • Top of funnel increases • Increased conversion • Claims Data Collection • Detailed follow-up • Email drafting • Internal Help Desk • Self service • Increased employee efficiency • Agent Support • Increased employee efficiency • Improved customer satisfaction • Policy Pricing • Optimize against expected outcomes • Leverage third party data • Hold vs Sell Optimization • Leverage additional information for hold/sell • Advertising Expenses • Optimized targeting • Increased yield models • Customer Insights • Increase NPS score • Proactively identify churn • Product Development • Identify new behaviour patterns • Claims Processing • Automate basic tasks • Reduce human error • Fraud Detection • Improve fraud detection • Increased staff efficiency • Client Onboarding • Automate basic tasks • Manage intake of documents • Documentation processing • Automate basic tasks • Email drafting • AR / AP Automation • Automate report creation • Automate collection activities
  • 16.
    16 © Copyright Publicis.Sapient| Confidential WHAT YOU KNOW 1st Party Data WHAT THEY SAY Social Behavior WHAT THEY DO Consumer Behavior THEIR OPINIONS Survey Data Step 2: Define Your Data Needs WHY
  • 17.
    C O PY R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 17 Step 3: Define the Services Text Generation Query Parser Results Post-Processor Indexing Structured Data Access Ontology Creation Ontology Evolution Facets/Filters Relevance-based ResultsRanking UserEntitlements Check Tagging/Metadata Extraction Content Storage Concepts/Entities Extraction Relations/Fac t Extraction Heuristic Inference Deductive SERVICES API (KNOWLEDGE ENGINE ABSTRACTION LAYER) Concepts Disambiguation Free-Text Search Contextual Search Graph Search Reasoning / Explanation Query Formulation Query Expansion NLP / NLU Chat Agents Voice & Tone Mobile Framework ASR / TTS USER EXPERIENCE FRAMEWORK Reporting / Portal Email Text Interfaces Audio Capture TEXT & AUDIO INTERFACES Audio Output REST / SOAP APIs Managed File Transfer Security Logging Caching Monitoring UserEntitlements REASONING ENGINES INTEGRATION & CORE SERVICESKNOWLEDGE BASE MGMT DATA & CONTENT INGESTION DATA & CONTENT DELIVERY Topic Modelling, Query Exapansion, NLP/NLU, Ontology Indexes, Structured Data Triples Knowledge Base, Reasoning Natural Language Translation Deep Learning Engines Constrained Conditional Models Ontology Neural Networks KNOWLEDGE SERVICES MACHINE LEARNING
  • 18.
    C O PY R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 18 Step 4: Align to Technology Platforms Text Generation Query Parser Results Post-Processor Indexing Structured Data Access Ontology Creation Ontology Evolution Facets/Filters Relevance-based ResultsRanking UserEntitlements Check Tagging/Metadata Extraction Content Storage Concepts/Entities Extraction Relations/Fac t Extraction Heuristic Inference Deductive SERVICES API (KNOWLEDGE ENGINE ABSTRACTION LAYER) Concepts Disambiguation Free-Text Search Contextual Search Graph Search Reasoning / Explanation Query Formulation Query Expansion NLP / NLU Chat Agents Voice & Tone Mobile Framework ASR / TTS USER EXPERIENCE FRAMEWORK Reporting / Portal Email Text Interfaces Audio Capture TEXT & AUDIO INTERFACES Audio Output REST / SOAP APIs Managed File Transfer Security Logging Caching Monitoring UserEntitlements REASONING ENGINES INTEGRATION & CORE SERVICESKNOWLEDGE BASE MGMT DATA & CONTENT INGESTION DATA & CONTENT DELIVERY Topic Modelling, Query Exapansion, NLP/NLU, Ontology Indexes, Structured Data Triples Knowledge Base, Reasoning Natural Language Translation Deep Learning Engines Constrained Conditional Models Ontology Neural Networks KNOWLEDGE SERVICES MACHINE LEARNING Bespoke & Point Solutions Semantic AI Natural Language & Vision Platforms Align with Internal Arch Deep Learnin g ML EngineKnowledge Engines (ML, Semantic, NLU)
  • 19.
    C O PY R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 19 KNOWLEDGE LEVERAGE SCALABILITY EXPERIENC E DESIGN SECURITYLEARNING Step 5: Execute Real-World Solutions BUILD AS AN INTEGRATED PLATFORM NOT AS POINT SOLUTIONS
  • 20.
    C O PY R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 20 Step 6: Incorporate Experience Design Happy Path
  • 21.
    C O PY R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 21 Step 6: Incorporate Experience Design Happy Path Sad Facts Changed Mind Make No Sense Unexpected Question Small Talk
  • 22.
    C O PY R I G H T S A P I E N T R A Z O R F I S H | C O N F I D E N T I A L 22 Step 7: Iterate DATA NEEDS HUMAN BEHAVIOUR TECHNOLOGY SHIFTS LEARN FROM FAILURE FIND NEW OPPORTUNITIES

Editor's Notes

  • #3 Key theme: The intersection of Data and AI isn’t just another new technology wave, but rather a new business paradigm.
  • #4 Key theme: Things are moving much faster than people think. Not that it won’t be possible, but playing catch-up to your competitors will be a much bigger hurdle than it has been in the past.
  • #5 Key theme: Things are moving much faster than people think. Proof points. Hit the 25% of jobs impacted by 2019 hard – three years is close enough to be scary…
  • #6 Key theme: There will be some real winners. And the $1T per five years is something that Yuri expects to continue. Yuri Milner is the CEO of DST Global, an early investor in Facebook, Twitter, and Whatapp.
  • #7 There will also be a number of companies that don’t successfully transform. Richard Foster, a lecturer at the Yale School of Management, hasfound that the average lifespan of an S&P company dropped from 67 years in the 1920s to 15 years today. Foster also found that on average an S&P company is now being replaced every two weeks, and estimates that 75 percent of the S&P 500 firms will be replaced by new firms by 2027.
  • #8 Good educational video…
  • #9 Paving the cow-path will not get the job done. People need to reassess their entire business models in light of what data and AI make possible today.
  • #11 These three areas of transformation are a clean way to think about potential changes for any firm’s business processes.
  • #12 Common use cases: Service desks, virtual assistants, intelligent search
  • #17 Imagine the possibility to understand individuals not only by the medical data on a ‘patient’, But by connecting all the “1st party data” you know from when they interact with your sites, content, media… to create a 360o view of “the person”. Then if we add in their social behavior – what they SAY, And add in consumer behavior – what they DO. And layer on causal intelligence to look at why, and what it means. We can start to see individual data-driven journeys. And imagine if we were able to leverage those at scale. The Behavioral Insights can Drive Engagement Model
  • #18 These three areas of transformation are a clean way to think about potential changes for any firm’s business processes.
  • #19 These three areas of transformation are a clean way to think about potential changes for any firm’s business processes.