Cognitive Business

This content included for educational purposes. 2
• Lawrence Mills Davis is founder and managing director of
Project10X, a research consultancy known for forward-looking
industry studies; multi-company innovation and market
development programs; and business solution strategy
consulting. Mills brings 30 years experience as an industry
analyst, business consultant, computer scientist, and
entrepreneur. He is the author of more than 50 reports,
whitepapers, articles, and industry studies.
• Mills researches artificial intelligence technologies and their
applications across industries, including cognitive computing,
machine learning (ML), deep learning (DL), predictive analytics,
symbolic AI reasoning, expert systems (ES), natural language
processing (NLP), conversational UI, intelligent assistance (IA),
and robotic process automation (RPA), and autonomous multi-
agent systems.
• For clients seeking to exploit transformative opportunities
presented by the rapidly evolving capabilities of artificial
intelligence, Mills brings a depth and breadth of expertise to help
leaders realize their goals. More than narrow specialization, he
brings perspective that combines understanding of business,
technology, and creativity. Mills fills roles that include industry
research, venture development, and solution envisioning.
Lawrence Mills Davis
Managing Director
Project10X
mdavis@project10x.com
202-667-6400

This content included for educational purposes.
Introduction
3
Cognitive business — Insight, engagement, acceleration &
transformation  
Enterprise AI creates opportunities to transform business, impact
performance significantly, and increase efficiencies:
• Insight generation is the production of an accurate and deep
intuitive understanding of a person or thing. Big data and
cognitive analytics extract previously unknown insight from
structured and unstructured data to both identify and act on
the opportunities presented.
• Customer engagement is the use of AI, information, analytics,
and communications technology to attract, involve, and support
someone’s interest, attention, interaction, and participation
towards some end. We explore application of AI to
conversational interfaces, bots, assistants, and virtual agents,
and to precision marketing, sales, and services.
• Business acceleration is the automation of knowledge
generation that drives cost savings, competitive advantage, and
new business lines through smarter deployment of resources.
Optimization is the action of making the best or most effective
use of a situation or resource. We explore cognitive enterprise,
cognitive platforms, intelligent automation, and intelligent
ecosystems.
• Enterprise transformation is the change associated with the
application of digital technologies and artificial intelligence to
all aspects of the business, its ecosystem, and human society.
We explore types of innovation and cycle time, perspectives on
cognitive enterprise transformation from selected industry
leaders, trends towards exponential value generation, and
implications enterprise AI for future skills.

1. Insight generation
2. Customer engagement
3. Business acceleration
4. Enterprise transformation
SECTIONS

Enterprise connected intelligence use cases — business and IT
Source: WIPRO
• Digital Virtual Agents—Enhanced user experience with
capabilibes like speech recognibon, natural language
understanding;  
E.g. Collaborabve Agents, Customer Support/Experience,
DIY Support.
• PredicGve Systems—Extracbng meaning from different
forms of data, using tools and techniques – to discover
paherns, predict future outcomes and trends;  
E.g. Recommender Systems, Anbcipatory Systems,
Automated Scenario Modeling.
• CogniGve Process AutomaGon—Cognibve Process
Automabon is defined and executed based on a loose set
of instrucbons or logic. These instrucbons are largely
machine-learnt, evolve conbnuously and can be user-
defined as well;  
E.g. Automated Problem Resolubon, Sojware Release
Automabon, Modal Interacbons and Experience
Management.
• Visual CompuGng ApplicaGons—Visual compubng
applicabons that can acquire, analyze and help synthesize
realisbc interacbve interfaces and idenbfy paherns; 
E.g. Dynamic Pahern Clustering, Computer Vision.
• Knowledge VirtualizaGon—System that can curate
knowledge by using AI techniques. They rely on usage of
expert knowledge databases to arrive at decisions;  
E.g. Diagnosbc Experts, Advisory Systems, Natural
Language Generabon.
• RoboGcs and Drones—Robobc automabon is powered by a
repebbve set of instrucbons. These instrucbons are mostly
defined by the user and somebmes machine-learnt. They
can be fed into the system by analyzing repebbve paherns;  
E.g. Smart Drones, Brain-Controlled Robobcs.
5

AI TECHNOLOGY EXAMPLEENTERPRISESOLUTION
Computer vision
Acquiring, processing, analyzing and
understanding images
Video analytics integrated with surveillance cameras provides situational awareness of
business operations, delivering insights about risk, safety and security. In retail, video
analytics can be used to gain insights into shopper behaviors effectively and
systematically.
Audio processing
Identifying, recognizing and analyzing sounds and
speech
Speech recognition technologies integrated into call centers automate the
identification of callers.
Sensor processing
Processing and analyzing information from
sensors other than cameras and microphones
In an agricultural setting, sensors in the field can be integrated with software to
deliver “precision agriculture”— sensing and communicating status about
temperature, humidity, etc., enabling more precise care for crops.
Natural language processing
Understanding and generating language in
spoken and/or written form
Personal assistants on consumer smart phones provide guidance and services using
natural language. Increasingly, search capabilities include the ability to understand
the meaning of what a person is saying, not just recognizing key words or doing
statistical retrieval.
Knowledge representation
Depicting and communicating knowledge to
facilitate inference and decision making
Knowledge-based tools provide the capability to link a particular search or piece of
content to other relevant content on the web. This is done by tagging all content
and then mapping it to a larger representation of knowledge. For example, a search
for “Da Vinci” will link one to particular paintings and creations, as well as to Italy,
to the Renaissance, and so forth.
Inference engines
Deriving answers from a static knowledge base
such as business rules
Solutions can apply rules to make automated loan approval or credit decisions, or
granting of visas. Such capabilities can deliver accurate decisions in a fraction of the
time of manual decision making.
Expert systems
Reasoning with rules, algorithms and
information available in its knowledge base
Medical diagnostics as well as legal research can be significantly aided by the ability
of expert systems to sift through millions of data sources, synthesize information
and present it to a user.
Machine learning
Altering the decision process based on experience
Software tools and personal agents can learn from users to improve
productivity—for example, by sorting email, then extracting calendar entries
and action items.
Source: Accenture
Examples of AI technologies
integrated into enterprise
solutions
6This content included for educational purposes.

• What is insight generation?
• Analytics continuum
• Turning structured and unstructured data into actionable insights
• Types of (big) data used in predictive marketing
• Types of analytics used in predictive marketing
• Machine learning to optimize targeting — with example
• How machine learning predictive analytics works
• 40+ machine learning and predictive analytics use cases across
industries
• Sapient cognitive analytics — Cosmos, Idiom & Luminoso
Overview of 
insight generation

in·sight gen·er·a·tion
/ˈinˌsīt/ /ˌjenəˈrāSH(ə)n/
The production of an accurate and deep intuitive understanding of a
person or thing.
Big data and cognitive analytics uncover previously hidden patterns
and relationships from structured and unstructured data to both
identify and act on the opportunities presented for innovation,
growth, diversification, and efficiencies.
What is insight
generation?
!

Analytics continuum and stages
Information  
Foundation
Descriptive
Diagnostic
Predictive
Prescriptive
Cognitive
RDBMS
Information 
Integration
ECM
Big Data 
Platforms
Standard 
Reports
Ad Hoc 
Reports
Drilldown 
Query
Alerts
Statistical 
Analyses
Forecasting & 
Extrapolation
Predictive 
Modeling
Rules
Simulation/ 
Optimization
Natural Language Learning
Reasoning/ 
Explanation
Recommendation
Prediction
Data DecisionInsight
Context specific use
What's the best that can happen?
What will happen next?
Why is this happening?
What if trends continue?
Where exactly is the problem?
What actions are needed?
What happened?
Who? What? When? How often? How much?
How do we know?
10

Journey from analytics to cognitive computing — capture increasing value  
through outcome-driven, actionable insights
Source: HfS
11

| INSIGHT GENERATION
MACHINE LEARNING
PURCHASETV
SMART
PHONE
DIGITAL
An Operating System Turning Data into InsightsSTRUCTURED
DATA
TV
MOBILE
DEVICES
SOCIAL
UNSTRUCTURED  
DATA
| |
ACTIONABLE
INSIGHTS
80% +
of all data is
unstructured
12
Source: Publicis•Sapient

From text to actionable insights
Source: SearchBlox
13

What types of (big)data are used in predictive marketing?
First-party data
▪ Internal data includes anything sitting in a
data warehouse, CRM system, or other
sources that have not been integrated into
your marketing database.
▪ Examples of internal data include customer
service records, transactional data, credit
card purchases, or contact information
provider by the customer.
Second-party data
▪ Internal data purchased from a business or
traded for that includes anything sitting in a
data warehouse, CRM system, or other
sources that have not been integrated into
your marketing database.
Third-party data
▪ External data is available for purchase by
data providers who source and aggregate
the data into applicable sets that can be
applied to first party databases.
▪ 3rd party data enhances targeted marketing
campaigns, because it provides hundreds of
detail elements that no consumer would fill
out in a single form.
▪ With only a few first-party data elements,
third-party data sets can be appended to
correct and fill in missing elements such as
email addresses, phone numbers, lifestyles,
demographics, purchase indicators and more
to strengthen customer insights.
In-market signals
▪ Today’s always-on and connected consumer
leaves a digital footprint indicating in-market
purchase signals. Advancements in technology
have made it possible to match a consumer’s
mobile ID to a piece of PII (Personal Identifiable
Information), which can be matched to social
IDs and IP addresses to determine search data.
▪ Matching offline and online data establishes rich
consumer profiles and access realtime digital
behavioral data indicating life events and
purchase intent.
▪ For instance, social signals are created when
people post to social networks about “Moving
to Denver”, “Taking a family vacation to
Orlando”, or “Looking for recommendations on
a new car.” Also, search data created when
consumers research new cars or furniture or
browse on e-commerce sites can be used as
indicators of life events and intent.
14

What types of analytics are used in predictive marketing?
Descriptive
▪ Descriptive Analytics give hindsight or insight into
the past. They use data aggregation and data mining
techniques to provide insight into the past and
answer: “What has happened?”
▪ Descriptive statistics help understand raw data at an
aggregate level, to learn what is going on, and to
summarize and describe different aspects of the
business. Usually, the underlying data is a count, or
aggregate of a filtered column of data to which basic
math is applied like sums, averages, and percent
changes.
▪ Descriptive analytics allow learning from past
behaviors, and how they might influence future
outcomes. The past, which can be any point of time
that an event has occurred, whether it is one minute
ago, or one year ago. Descriptive statistics show
things like, total stock in inventory as of a point in
time, average dollars spent per customer, and year
over year change in sales.
▪ Typical outputs include reports that provide
historical insights regarding the company’s
production, financials, operations, sales, finance,
inventory and customers.
Predictive
▪ Predictive Analytics give foresight. They use
statistical models and forecast techniques to
understand the future and answer: “What could
happen?”
▪ Predictive analysis is used when there is a need to
estimate something about the future, or to fill in the
information gaps.
▪ Statistical algorithms combine historical data found
in ERP, CRM, HR and POS systems, and may enhance
these with information from public records and 3rd-
party data sources to identify patterns in the data,
and apply statistical models and algorithms to
capture relationships between various data sets.
▪ They estimate the likelihood of a future outcome
based on probabilities, hence with some
uncertainty.
▪ Examples include forecasting customer behavior
and purchasing patterns such as what items
customers will purchase together, to identifying
trends in sales activities, to forecasting demand for
inputs from the supply chain, operations and
inventory based upon a myriad of variables.
Prescriptive
▪ Prescriptive Analytics: advise on possible outcomes
and next best actions. They use optimization and
simulation algorithms to advice on possible outcomes
and answer: “What should we do?”
▪ Prescriptive analytics seek to quantify the effect of
future decision alternatives in order to advise on
possible outcomes before the decisions are actually
made.
▪ The goal is to predict not only what will happen, but
also why it will happen, providing recommendations
regarding next best actions that will take advantage
of the predictions.
▪ Prescriptive analytics combine multiple techniques
and tools such as business rules, algorithms, machine
learning, and computational modeling procedures,
and apply these against input from many different
data sets including historical and transactional data,
real-time data feeds, and big data.
▪ Examples of prescriptive analytics applications
include simulation and optimization of production,
scheduling, and inventory in a supply chain to ensure
delivery of the right products at the right time while
optimizing customer experience.
15

DATA LOGIC EXPERIENCES MEASUREMENT
Marketing Database
Optimizations Targeting Rules
3rd Party
Social
Product Data
Interactions
Transactions
Customer
Media
Performance
Sales Data
360° Customer
Profile
Data
Warehouse
UNKNOWN
DMP
KNOWN
Campaign
Management
Personalization
Content /
Campaign
Testing
Rules
Content & Asset
Management
Product Info
Management
Transaction
Engine
Order
Management
CRM
Tag Mgmt
Custom  
Services
EngagementAnalytics
Modeling
Multi-channel
Attribution
Customer  
Value
Business KPIs
& Scorecards
Experience
Level
Optimization
Monetization
Call Center
Retail/POS
Sales
Email
Social
Mobile
Web
Online Storefront
IoT
Media
Direct Mail
DataServiceLayer
ExperienceServiceLayer
| INSIGHT GENERATION: Machine Learning to Optimize Targeting
16

Basic Analy+cs
• Counts and Averages
• Data Preprocessing
• SQL Queries
• Human Scale
• Hand Crajed
Advanced Analy+cs
• Discovering Paherns
• Making Predicbons
• Mulbvariate Queries
• Machine Scale
• Data Driven
Image source: http://personalexcellence.co/blog/ideal--beauty/
Sample predictive
marketing data
17

Size of Network
Lifestyle
ZIPcode
Costal vs Inland Marital status
Generation
Family Size
GenderIncomeLevel
Competitors
Age
Revenue Size
Life Stages
Education
Location
Sector
Industry
Legal status
City
Loyalty and card activity
Basic personal data
Sample predictive
marketing data
18

Size of Network
Lifestyle
ZIPcode
Generation
Family Size
Gender
IncomeLevel
Competitors
Age
Revenue SizeLife Stages
Education
Location
Sector
Industry
Legal status
City
Size of Network
Subscriptions
Date on Site
Wish List
Deposits/Withdrawals
Device Usage
Following
Followers
Likes
Sequence of visits
Time/Day log in
Time spent on siteVideos Viewed
Photos liked
Check-ins
Number of Apps on Device
App usage duration
Number of Hashtags used
Frequency of Search
History of Hashtags
Search Strings entered
Purchase History
Time spent on page
In market signals 
and social media  
acbvity data
Sample predictive
marketing data
19

Sentiment
Lifestyle
ZIPcode
Generation
Family Size
Gender
IncomeLevel
Competitors
Age
Revenue SizeLife Stages
Education
Location
Sector
Industry
Legal status
City
Size of Network
Subscriptions
Date on Site
Wish List
Deposits/Withdrawals
Device Usage
Following
Followers
Likes
Sequence of visits
Time/Day log in
Time spent on siteVideos Viewed
Photos liked
Check-ins
Number of Apps on Device
App usage duration
Number of Hashtags used
Frequency of Search
History of Hashtags
Search Strings entered
Purchase History
Time spent on page
Tone
Euphemisms
Hedonism
Extroversion
Face Recognition
Openess
Colloquialism
Reasoning Strategies
Language Modeling
Dialog
Latent Semantic Analysis
Linguistics
Image Tags
Question Analysis
Self-transcendent
Affective Status
Phonemes
Intent
Insights derived from 
cognibve analybcs
Sample predictive
marketing data
20

Gaining customer insight through predictive analytics
Prediction data output includes:
• Buyer identification
• Enhanced consumer profiles,
behaviors, and market signals
• Specific life event knowledge, facts,
and statistical probabilities on which
the prediction is based
• Type of product or service transaction
• Estimated transaction size
• Timing of the close
• Confidence index (probability)
Steps in the predictive process:
• Author knowledge about selected
life events
• Develop predictive models using
statistical machine learning based
on 1st, 2nd, and 3rd party data
• Make predictions of demand and
personalization needs
• Handle queries from and deliver
predictions to activation systems
21

How does a machine learning predictive analytics application work?
The mortgage life event predictor POC will deliver significant business
value for the bank partner:
▪ Life event predictive marketing can drive quantum improvements in
marketing performance compared to demographic approaches.
▪ At any time about 1/3 of bank customers are expecting a life event.
Bank customers buy or shed products within 30 days of a major life
event. Consumers are 40% more likely to buy a financial product
around a life event.
▪ A bank customer typically will have on average 4-to-5 life events that
need managing and supporting every 24-month cycle. This means a
running opportunity exists to sell at least 4 products per banking
household on a rolling 2 year basis.
▪ Life event prediction delivers high value prospects, who are more
likely to engage with relevant offers.
▪ Predictive marketing provides early notification and first mover
advantage in relationship banking.
▪ Enhanced consumer profiles enable customer-centric segmentation
and deep personalization of communications and interactions.
▪ Life event behavior based customer insight has been shown to
improve response rates by 2-to-10 times, improve customer lifetime
value by 25%, increase loyalty by 30%, and increase retention by 30%.
▪ Life event targeting has been shown to lift conversion rates by 40%,
reduce default rates by 30%, and lower customer acquisition costs by
up to 50%.
22
Train / Retrain
(SML)
Model /
Algorithm
Predict
(Statistical + Causal)
Notification
(API)
Knowledge authoring
& curation
Historical
1st, 2nd & 3rd
party data
Life Event
Knowledge Graph
New
1st, 2nd & 3rd
party data
Recommendation  
& Explanation
Evaluation
1 CONFIGURE AI PLATFORM CAPABILITIES
2 TRAIN PLATFORM 3 MAKE PREDICTIONS 4 ACTIVATE NEXT BEST ACTIONS
5 IMPROVE PERFORMANCE WITH USE AND SCALE
Next
Best
Actions

• Go-to-market excellence requires bringing “data sophistication” to all six
facets of competitiveness:
- Market intelligence and strategic priorities
- Product development and portfolio management
- Marketing and communication
- Sales platform management
- Performance monitoring
- Organizational enablers such as recruiting, compensation, and training.
• The right kind of platform is critical to tap the power of advanced analytics.
Organizations need a flexible platform that centralizes their data, and
allows it to be analyzed from any perspective.
• Also essential are the ability to incorporate external and unstructured data
streams, a robust user interface adaptable to varying management needs,
and open architecture that leaves room for future innovations.
Bringing data
sophistication to  
six facets of
competitiveness

Example marketing & sales predictor applications (1 of 2)
• PredicGng LifeGme Value (LTV)— If you can predict the
characterisbcs of high LTV customers, this supports
customer segmentabon, idenbfies upsell opportunibes and
supports other markebng inibabves. Usage can be both an
online algorithm and a stabc report showing the
characterisbcs of high LTV customers
• Wallet share esGmaGon — Working out the proporbon of a
customer's spend in a category accrues to a company allows
that company to idenbfy up-sell and cross-sell
opportunibes. Usage can be both an online algorithm and a
stabc report showing the characterisbcs of low wallet share
customers
• Churn — Working out the characterisbcs of churners allows
a company to product adjustments and an online algorithm
allows them to reach out to churners. Usage can be both an
online algorithm and a stabsbcal report showing the
characterisbcs of likely churners
• Customer segmentaGon — If you can understand
qualitabvely different customer groups, then we can give
them different treatments (perhaps even by different groups
in the company). Focus is to answer quesbons like: what
makes people buy, stop buying etc. Usage: can be guidance
• Product mix — What mix of products offers the lowest
churn? eg. Would giving a combined policy discount for
home + auto result in low churn? Usage: online algorithm
and stabc reporting
24

Example marketing & sales predictor applications (2 of 2)
• Cross-selling & RecommendaGon algorithms — Given
a customer's past browsing history, purchase history
and other characterisbcs, what are they likely to want
to purchase in the future. Usage can be an online
algorithm
• Up-selling — Given a customer's characterisbcs, what
is the likelihood that they'll upgrade in the future?
Usage can be online algorithm and stabc report
• Channel opGmizaGon — What is the opbmal way to
reach a customer with certain characterisbcs? Usage
can be an online algorithm and stabc report
• Discount targeGng — What is the probability of
inducing the desired behavior with a discount? Usage:
online algorithm and stabc report
• CalculaGng the right price for diﬀerent keywords/ad
slots — What is the reacbvabon likelihood for a given
customer? Usage can be an online algorithm and stabc
report
• Adword opGmizaGon and ad buying — What is the
right price for diﬀerent keywords/ad slots?
25

9 machine learning use cases
Supervised Learning
Predict credit worthiness of credit card
holders: Build a machine learning
model to look for delinquency
attributes by providing it with data on
delinquent and non-delinquent
customers
Predict patient readmission rates: Build
a regression model by providing data on
the patients' treatment regime and
readmissions to show variables that
best correlate with readmissions
Analyze products customers buy
together: Build a supervised learning
model to identify frequent item sets
and association rules from transactional
data
Unsupervised Learning
Segment customers by Survey
prospects and customers to develop
multiple segments using clustering
behavioral characteristics
Categorize MRI data by normal or
abnormal images: Use deep learning
techniques to build a model that
learns different features of images to
recognize different patterns
Recommend products to customers
based on past purchases: Build a
collaborative filtering model based on
past purchases by "customers like
them"
Reinforcement Learning
Create a 'next best offer' model for the
call center group: Build a predictive
model that learns over time as users
accept or reject offers made by the
sales staff
Allocate scarce medical resources to
handle different types of ER cases:
Build a Markov Decision Process that
learns treatment strategies for each
type of ER case
Reduce excess stock with dynamic
pricing: Build a dynamic pricing model
that adjusts the price based on
customer response to offers
Banking
Healthcare
Retail
26

Machine learning
applications across
industries
Source: Forbes

PUBLICIS.SAPIENT  
COGNITIVE ANALYTICS
Source: Publicis•SapientThis content included for educational purposes.

AUDIENCE INTELLIGENCE
Build deeper relationship with
your audiences with One-to-One
Cognitive Media Engagements.
• Consumer 360 Data Strategy
• DMP Activation Strategy
• Media Campaign Automation
• Media Optimization Maturity
• One-to-One Targeting Strategy
Engage your customer with
rich, dynamic and relevant
brand experiences.
• CRM Data onboarding
• Customer Data Normalization
• Persona Activation
• Customer Journey Mapping
• Customer Lifetime Value
Activate cross-channel brand
experiences that drive  
business impacts.
• Audience Segmentation
• Programmatic Targeting
• CRM & 3rd Party Data Alignment
• Probabilistic Matching
• Cross-Channel Activation
COGNITIVE AUTOMATION
Automate smarter experiences
with artificial intelligence to
disrupt your industry.
• Fraud & Anomaly Detection
• Media Buying Automation
• Programmatic Creative
• Relevance Down the Path to Purchase
MARKETING ANALYTICS
Gather, analyze and act upon your
customer data in real time and
across all marketing
• Audience Graph Analysis
• Social Sentiments Intelligence
• Predictive & Prescriptive Analytics
• Regency and frequency Reporting
• Cross-channel attribution
COSMOS™ Cognitive Marketing Intelligence Platform
COSMOS™
Cognitive Services APIs
COSMOS™
Audience Intelligence
Audience Network
COSMOS™
Attribution Analytics
COSMOS™
DOMO Integration APIs
COSMOS™
Universal Graph ID
COSMOS™
Consumer 360
COSMOS™
Social Sentiments Graph
BUSINESS IMPACTS
DEEP PERSONALIZATION
ONE-TO-ONE MARKETING
CONSUMER JOURNEY
AUDIENCE DATA
ALIGNMENT
SINGLE VIEW OF THE
CUSTOMER
AUDIENCE ACTIVATION
BID OPTIMIZATION ATTRIBUTION ANALYTICS
PROGRAMMATIC TAGS REAL TIME DECISIONNING
COGNITIVE MEDIA STRATEGY CONSUMER 360 GRAPH COGNITIVE AUTOMATION MARKETING ANALYTICS
| INSIGHT GENERATION: COSMOS™ Powered Media Solutions
Source: Publicis•Sapient This content included for educational purposes.

July 1 July 15 July 30
$10.00
$5.00
$0.00
$30.00
$20.00
$100.00
$90.00
$80.00
$70.00
$60.00
$50.00
$40.00
120
60
80
20
40
100
0
Cost per Action
Number of Attributes
+2.11 Visited a product page
+0.95 Is in the Atlanta DMA
+0.89 Saw an ad 7-14 days ago
-0.55 Is planning a trip
+0.60 Is reading the news
+0.51 Searched for luxury products
-0.65 On a mobile device
+0.46 Made a luxury retail purchase
+0.43 Played an online game
-0.45 Purchased sporting goods
-0.54 Made a non-luxury retail purchase
+0.41 Is in the Tampa DMA
-0.50 Watched a TV show online
+0.38 Clicked on an ad before
+0.31 Booked a flight in the last week
+0.29 Saw an ad 1-7 days ago
+0.25 Is in the Orlando DMA
-0.54 Is searching for an apartment
-0.21 Is in the Los Angeles DMA
-0.30 Saw an ad within the last hour
+0.37 Has clicked on an ad before
-0.31 Has seen 3+ ads already
+0.48 Is in the Houston DMA
-0.50 Watched a TV show online
-0.54 Made a non-luxury retail purchase
COSTPERCUMULATIVEACTION
CAMPAIGN TIMELINE
SIGNIFICANTMODELATTRIBUTES
RELEVANTATTRIBUTES  
“MICRO-MOMENTS”
| INSIGHT GENERATION: COSMOS Learns and Optimizes from Real-time
Micro-Moments
30

PROGRAMMATIC MEDIA
• Micro-Moment Targeting
• Bid Impression Value (BIV)
• eCPM Optimization
• RTB Optimizer
• Attributions
• Ad Creative Personalization
• Ad Serving Fraud Detection
• Micro-Moment Segmentation
• Lifetime Value (LTV)
• Propensity
• Recency, Frequency and Monetary
Value (RFM)
• Churn Prediction & Prevention
• Universal ID Sequencing
• Sentiments Signals
• Message Resonance
• Concept Expansion
• Face Detection
• Natural Language Classifier
• Speech to Text
• Text to Speech
• Language Translation
• Language Detection
• Sentiment Analysis
• Dialog
• Retrieve and Rank
• Image Link Extraction
• Tradeoff Analytics
• Entity Extraction
• Tone Analyzer
• Personality Insights
• Taxonomy
COGNITIVE APIs
• Intelligent Search
• Frequently Bought Together (FBT)
• Cross-Selling (item correlations)
• Sentiment and trend analysis
• Shipping cost and time estimation
• Logistics optimization
• Fraud detection and prevention
• Supply and demand analysis and forecast
• Wallet management and funding source
optimization
• Various scheduling and optimal resource
allocation
• Attributions
• Content Personalization
• Customer Lifetime Value C(LTV)
• Customer Propensity
• Recency, Frequency and Monetary Value
(RFM)
• Churn Prediction & Prevention
COGNITIVE COMMERCE
• Programmatic Creative | DCO
• Content Personalization
• Micro-Moment Segmentation
• Universal ID Sequencing
• Text Mining
• Sentiments Signals
• Cross-Screen Equalizer
• Auto-suggest Indexer
UNIFIED EXPERIENCE
Intelligence
AMPLIFY CONSUMER 360
COSMOS Artificial Neuro Network
• Author Extraction
• Concept Tagging
• Relationship Extraction
• Concept Insights
• Question & Answer
• Feed Detection
• Keyword Extraction
• Visual Recognition
• Image Tagging
• Text Extraction
| INSIGHT GENERATION: COSMOS™ Cognitive Library

Sentiment Trends
Positive sentiment is also highly emotive
and is closely associated with artist
fandom. Fans proclaim love and
excitement around iHeartRadio songs and
events, and interact with artist-centric
iHeartRadio social content.
Association scores range on scale from -1.00 to
1.00. These are extremes which represent the
weakest and the strongest possible relationships:
An association score of 100 represents the
relationship between a concept and itself, while -100
is the relationship between a concept and the most
unrelated other concept within the same data set. A
score of association score of around 0 represents
how much we would expect two concepts to be
discussed at the same time as a result of random
chance.
Above: Top concepts  
(tool-defined based solely on
phrase occurrence)
associated with positive
sentiment.
Right: Examples of
stereotypically emotive fans
currently listening to &
enjoying a song (left),
anticipating an event (right).
| INSIGHT GENERATION

To Create a Single Point of Truth for Each Consumer
To Eliminate Siloed
Digital Experiences
To Empower Cross Channel and Device
Personalize Experiences

To Connect Advertising,
Customer Service, Marketing, and CRM Tools
To move beyond rule based optimization
Creating 1 to 1 Optimization
To Enable Consumer Centric Marketing
WHY WE BUILT COSMOS?

34

| IDIOM CREATES THE SINGLE VIEW OF THE CONSUMER…
SARAH
Age 26 
Manhattan 
Account Exec
$75K
smgOS allows for greater insight into a brand’s “true” customer and/or
opportunity by unveiling behavioral truths
35Source: Publicis•Sapient

NOBODY HAS EVER BROUGHT THIS RANGE OF DATA TOGETHER BEFORE
• Multi-Client Handling
• Export
BDE & SAFE HAVEN
• Full panel data for search and
web
DIGITAL
• Co-developing exclusive
clustering methodology to
provide TV data at 30 HH
cluster level
TV
• Extensive data rights review
unveiling new insights for
potential partners
MOBILE
• “Reimagining our data in a
way its never been sold”
PURCHASE
36

38
COSMOS PROVIDES POWERFUL TOOLS FOR MARKETERS TO UNDERSTAND BEHAVIOR

Concept Cloud
39
At first glance, positive clusters
that jump out are the artist
names and birthdays. Negative
clusters appear associated with
news of a celebrity death, as
well as songs / concerts.
Mentions of iHeartMedia were
highly neutral and tended to
originate from industry & media
sources as opposed to consumers, so
we instead analyzed the more
consumer-driven, emotive body of
conversation around iHeartRadio,
keywording specifically to source
positive and negative themes (see
slide notes).
Concept cloud data, as well as data
used throughout this report, is drawn
from Twitter and Facebook mentions
of iHeartRadio from
1/1/2016-6/27/2016.

Sentiment trends
• Negative sentiment is highly
emotive, and often in
response to sad/provocative/
etc. iHeartRadio music and
content. People aren’t upset
with iHeart- they’re upset in
tandem with iHeart about
the emotional content iHeart
publishes, or about songs/
albums/artists.
Association scores range on scale from
-1.00 to 1.00. These are extremes which
represent the weakest and the strongest
possible relationships: An association
score of 100 represents the relationship
between a concept and itself, while -100 is
the relationship between a concept and
the most unrelated other concept within
the same data set. A score of association
score of around 0 represents how much
we would expect two concepts to be
discussed at the same time as a result of
random chance.
Above: Top concepts (tool-
defined based solely on phrase
occurrence) associated with
negative sentiment.
Right: Top concepts associated
with the more nuanced ‘sad,’
which is largely associated with
a celebrity death covered by
iHeartMedia in social.Source: Publicis•Sapient
40

• Customer engagement
• Conversational interface
• Bots
• Sapient AI platform for chatbots, assistants, and precision marketing
• Cognitive marketing, sales, and servicesOverview of 
Customer Engagement

cus·tom·er en·gage·ment
/ˈkəstəmər//inˈɡājmənt,enˈɡājmənt/
The use of AI, information, analytics and
communications technologies to attract,
involve, and support someone's interest,
attention, interaction, and participation
towards some end.
For example, using intelligent agents and
avatars to deliver hyper-personalization at
scale through all channels, including smarter,
more relevant insights and contextual
recommendations to amplify end-user
experience.
43

Our expectations have evolved. The era of consumer and enterprise
conversational computing is dawning.
Speech Enabled Devices
Virtual Assistants
Smart, Speech-Enabled Sites
Messaging & Social Media
44

We Amazon the diapersWe Netflix the showWe Uber the car
We Spotify the Song We Yelp the restaurant We Google the symptoms 45

“‘Conversational AI-first’ will supersede  
‘cloud-first, mobile-first’ as the most important,
high-level imperative for the next 10 years.”

CONVERSATIONAL
INTERFACE
48
CONVERSATIONAL INTERFACE
• Communication enabled by natural
language involving:
- Multiple contributions
- Coherent interaction
- More than one participant
• Multiple interaction modalities:
- Input: Speech, typing, writing, pictures,
gesture
- Output: Speech, text, graphical display/
presentation, animated face/body

What is involved in
conversational UI?
• Understanding:
-What does a person say?
‣ Identify words & other entities
from input signals
‣ “Please close the window”
-What does the speech, image or
gesture mean?
‣ Identify semantic content
‣ Request ( subject: close ( object:
window))
-What are the speaker’s intentions?
‣ Speaker requests an action in a
physical world

What is involved in conversational UI?
• Managing interaction:
- Internally representing the domain
- Identifying new information
- Deciding which action to perform given new information:
‣ “close window”, or “set thermostat” = physical action
‣ “what is weather outside?” = call the weather API
- Determining a response:
‣ “OK”, or “I can’t do it”
‣ Provide an answer
‣ Ask a clarification question
What is involved in
conversational UI?
• Managing interaction:
- Internally representing the
domain
- Identifying new information
- Deciding which action to
perform given new
information:
‣ “close window”, or “set
thermostat” = physical
action
‣ “what is weather outside?”
= call the weather API
- Determining a response:
‣ “OK”, or “I can’t do it”
‣ Provide an answer
‣ Ask a clarification question
50

• Access to knowledge and information
• E.g., to handle a request, “Please close the window”, the chatbot/assistant needs to know:
- There is a window
- Window currently is open
- Whether the window can or cannot be closed
What is involved in
conversational UI?
• Access to knowledge and
information
• E.g., to handle a request,
“Please close the window”,
the chatbot/assistant needs
to know:
- There is a window
- Window currently is open
- Whether the window can or
cannot be closed
51

• Producing language
- Deciding when to speak or otherwise respond
- Deciding what to say or display
‣ Choosing the appropriate meaning
- Deciding how to present information
‣ So partner understands it
‣ So expression seems natural
What is involved in
conversational UI?
• Producing language
- Deciding when to speak or
otherwise respond
- Deciding what to say or display
‣ Choosing the appropriate
meaning
- Deciding how to present
information
‣ So partner understands it
‣ So expression seems natural
52

When is a conversational interface useful?
53
• When hands-free interaction is needed:
- In-car interface
- In-field assistant system
- Command-and-control interface
- Language tutoring
- Immersive training
• When speaking is easier than typing and other mode of interaction:
- Voice as common interface across multiple platforms, devices and things
- Virtual assistant (Siri, Google Now, Cortana, etc.)
• When replacing or augmenting human agents:
- Voice interface for customer assistance and service provisioning
- Process and task automation
- Virtual assistance to improve capabilities, productivity, and efficiency of
knowledge workers
When is a conversational
interface useful?

BOTS

What is a bot?
• A bot is an autonomous
program on a network that can
interact with computer
systems or users to perform
tasks.
• Chatbots and virtual assistants
help customers and colleagues
perform tasks in increasingly
simpler and more effective
ways. Voice and text are the
most common modalities for
interacting with bots.

Bots are the new apps.
Conversations are the new UI.
AI is the protocol.
Messaging apps are the new browser.
56

| CUSTOMER ENGAGEMENT — Visions from science fiction cinema
STAR TREK (1966)
Natural language  
command and control
HAL “2001: A SPACE ODYSSEY” (1968)
Naturally conversing computer
HER (2013)
A virtual partner with natural
dialogue capabilities

JARVIS is not cool because of what Tony Stark can do with it, but because it is JARVIS

| CUSTOMER ENGAGEMENT — Evolution of intelligent agents
TOOL APPLIANCE CHATBOT ASSISTANT EXPERT SAVANT
TODAY 2018 2019-2025 2030+

TOOL APPLIANCE CHATBOT ASSISTANT EXPERT SAVANT
Tool requires detailed
procedural interaction
by user to perform a
sequence of steps to
accomplish function.
Chatbot is a conversational agent that
interacts with users using natural language
and AI. May have its own persona (avatar)
visualization. May act as virtual assistant.
Expert applies domain knowledge, deep
learning, task expertise, and legally defensible
reasoning to research, advise, and take actions
to solve complex problems requiring human-level
expertise.
Savant AI demonstrates far
better than normal human
capacities and abilities.
Assistant understands questions,
commands and intent; learns and
adapts to context, preferences, and
priorities; and marshals services and
information to accomplish tasks.
Appliance minimizes user steps to
specify and automate desired
function or service. User selects to
approve result or redirect.
Choosing the level of assistance
60

CHATBOT ASSISTANTAI CAPABILITIES
• Structured and unstructured
data ingest, cleansing and
curation
• Speech processing
• Knowledge acquisition
• Image processing
• Face and gesture recognition
• Emotion & sentiment
• Avatars
• Story and conversation
management
• Natural language
understanding
• Task and service orchestration
• Natural language generation
• Speech generation
• Visualization
• Presentation
• Speech and conversation
analytics
• Natural language processing
• Descriptive analytics
• Machine learning & deep
learning
•Predictive and
prescriptive analytics
•Knowledge management
•Semanticsearch
•Symbolic reasoning
• Question answering
• Advice & recommendation
• Next bestactions
• Expert assistance
• Taskplanning
• Command execution
• Data and service
provisioning
Capabilities today
61

Example: AI in hospitality apps, digital agents, and internet of things
CHAT CHATBOT ASSISTANT CONCIERGE BUTLER
SMART
FACILITY
Marriott
‘Mobile App’
Mobile app enables booking,
check-in/check-out, digital
room key, and requests to
staff before, during and after
the stay (via chat).
Radisson Blu
‘Edward’
Text-based virtual host
understands natural
language, handles digital
checkin, reports on hotel
amenities, gives directions
and tips, and receives
guest feedback and
complaints in a matter of
seconds via SMS.
Go Moment 
‘Ivy’
Smart texting platform for
hotels, powered by IBM
Watson AI, welcomes
guests, answers questions,
advises, integrates with
digital room key technology,
measures guest
satisfaction.
Hilton 
‘Connie’
Virtual concierge embodied
as NAO humanoid robot
that is approximately 23
inches tall. Connie answers
guest questions about hotel
amenities, local attractions
and dining options. It’s AI
uses IBM's Watson
machine-learning APIs, like
speech to text, text to
speech and its natural
language classifier. Connie
learns as it goes.
Starwood Aloft
‘Botlr’
Digital bellhop, or robotic
butler delivers amenities to
rooms. It knows the hotel
layout, is connected to
elevators, has avoidance
technology to not bump
into anything, and has a
touch screen for guests to
interact with it.
Marriott
M Beta
Hotel innovation incubator in
“live beta” From keyless entry
upon arrival, sensors
beacons enabling digital
experiences in the lobby,
fitness studio, meeting rooms,
cafe, and every corner of the
hotel. Infrastructure for rapid
prototyping, inviting guests to
test and give feedback in real-
time, ultimately shaping their
future hotel experience.
62

How intelligent chatbots work
Source: Inbenta
1. Captures data in real time
The intelligent chatbot captures the customer’s identity, attributes, and
engagement data, and any feedback the customer provides—all in real time.
For example, the chatbot determines:
• Date, time, physical location, and device information
• Whether the customer is on the web or a mobile app
• Whether the customer requested to engage with a chatbot or received a
proactive invitation
• Where the customer was on the website or mobile app when he or she
began the interaction with the chatbot
2. Uses internal data
Using data such as customer profile and preferences, value to the company,
location, industry, and amount of money spent in the past year gives the
chatbot more insights about the customer. This data is gathered from various
sources and is typically available in customer relationship management
(CRM) systems.
3. Combines data to predict customer intentions
The chatbot develops an understanding of what the customer wants/needs
by combining all the data signals. This helps make the conversation
contextual and more natural when the customer engages the chatbot.
4. Engages customers
Customers can invoke chatbots themselves when they need assistance, or
chatbots can proactively engage customers.
5. Understands what is said
The chatbot takes each message written or each utterance spoken and runs
it through natural language models to understand what the customer said.
This interaction is contextual and personalized to the customer. The chatbot
achieves this by leveraging information such as the web page the customer
was on when they engaged with the chatbot and their customer profile. For
example, if a customer is on a bank’s website looking at a page on mortgages
and asks the chatbot what the interest rate is, the chatbot will know the
customer is asking about the interest rate for mortgages.
6. Formulates a response
Once the chatbot understands the customer’s intent, the response-matching
algorithm determines the correct response and assembles it from knowledge
bases and CRM systems.
7. Determines follow-up actions
If the customer provides feedback that he or she is satisfied with the chatbot
response, the chatbot closes that intent and waits for a new intent. If the
customer requests the chatbot to help “pay my credit card bill,” for example,
the chatbot will determine the appropriate follow-up actions such as asking
the customer for a password and then completing the transaction.
63

Source: Inbenta
A chatbot should escalate to a live agent when:
1 2 3 4 5 6
The customer’s
request is not
understandable.
The customer
appears to be
annoyed or frustrated.
The customer’s
request cannot be
handled in self-service
(due to rules or
policies).
The customer’s
request is better
served by an agent
(e.g., conversion or
attrition).
It is a high-value
transaction and the
company wants a live
agent to close the
sales opportunity.
The customer explicitly
requests a human agent.
64

Digital advice is computer rendered guidance or recommendations
concerning future action.
For example, digital advisors for wealth management incorporate AI
technologies into their management processes – primarily through the
use of algorithms designed to optimize various elements of goal and risk
tolerance elicitation, to portfolio construction and asset allocation, to tax
management, to product selection and trade execution, to performance
monitoring and portfolio rebalancing.
Different digital advisors pursue different business models and
philosophies, and offer varying degrees of sophistication in services
provided. Also, the role of human involvement within digital advisors
varies.
https://www.blackrock.com/corporate/en-lm/literature/whitepaper/viewpoint-digital-investment-advice-september-2016.pdf
What is digital advice?

From voice commands & question answering to intelligent conversation
• Siri, Google Assistant, Cortana
and Alexa all essentially work
the same way — they recognize
and parse speech, classify
intent and execute commands.
• This framework works for
building a voice recognition
system that can interface with a
string of APIs, but it falls short if
you expect an intelligent
conversation.
• Intelligence requires more than
a great classifier. You need to
balance data, learning, memory,
computation and some
semblance of goals.
66
Bots learn to
converse
Source: Semantic Machines

Source: Inner Circle Guide to
Multichannel Customer Contact,
NewVoiceMedia, 2016.
Why chatbots now?
As the number of channels and
touchpoints multiply, customer
expectations continue to evolve
toward tailored, integrated
interactions and immediate answers
to their questions:
• 85% of consumers have used an
online channel for support
• 40% expect a response within the
hour
• 60% of consumers change
communication channels based on
where they are and what they’re
doing.

Source: Inbenta
Six ways enterprise chat bots and virtual assistants deliver value
1 2 3 4 5 6
Increase customer
self-service
engagement.
Improve customer
satisfaction ratings,
lower customer effort
scores, and increase
your Net Promoter
Score.
Automate routine
customer questions to
allow human agents to
focus on higher-value
interactions.
Deflect calls, email,
and chats to reduce
costs.
Minimize menial or
repetitive work for
agents.
Create a seamless
hand-off from self-
service virtual
assistance to a live
agent.
Maintain context of
previous interactions,
thus avoiding “starting
over.”
Reduce average
handling time by
suggesting responses
while the agent is
chatting with the
customer
Generate true “voice of
the customer” data
through the
conversations.
Mine agent interactions
to learn new customer
intents and agent
solutions.
Six ways enterprise chat bots and virtual assistants deliver value
68

Business value of enterprise chatbots across industries
Telecommunications
company Vodafone’s
virtual agent “Hani” is
an intelligent chatbot
that answers 80,000
questions per month
and deflects calls away
from the contact center
for 75 percent of the
customers it chats with.
Vodafone contact
center staff also use
the same technology to
access accurate, up-to-
date information on
Vodafone products
and services.
A leading global
airline created an
avatar to personify
their chatbot. The
chatbot serves as an
automated concierge,
providing customers
with instant, accurate
answers to their
questions about flight
status and baggage
rules. The chatbot
has helped the airline
reduce call and chat
volume by 40
percent.
Canadian Imperial
Bank of Commerce,
one of Canada’s
largest chartered
banks, introduced an
intelligent chatbot as
a virtual agent and
saw email volume
decrease by 50
percent immediately
at launch, and then
experienced another
23 percent drop
throughout the first
year. At the same
time, it reduced
phone calls by 25
percent.
A major health
insurance provider
improved the
experience for its 4
million members with
an intelligent chatbot
deployed as a virtual
agent. With the chatbot
answering 150,000
questions per month,
the company is saving
thousands of dollars in
contact center costs by
reducing calls to its
staff.
Canadian utility BC
Hydro improved
customer service and
satisfaction for its 4
million customers and
increased operational
efficiency by
deploying a chatbot
on its website. In the
first 11 months, the
chatbot answered
more than 720,000
questions with an
accuracy rate of 94
percent.
A major retailer
implemented an
intelligent chatbot to
deliver a phenomenal
guest experience,
answering 45,000
questions a month
about order status,
shipping, returns, and
other common areas of
interest. The chatbot
deflects informational
calls and email away
from staff by answering
97 percent of the
questions asked, with 96
percent accuracy.
Communications Travel Financial Services Healthcare Utilities Retail
69

The bot platform ecosystem
Nearly every large software company has announced
some sort of bot strategy in the last year. Here's a look
at a handful of leading platforms that developers might
use to send messages, interpret natural language. and
deploy bots, with the emerging bot-ecosystem giants
highlighted.

BOTS CAN HAVE
MASSIVE REACH
2.1+ BILLION ACTIVE USERS AND GROWING
900 M
25
170 M
26
2.7 M
27
275 M
28
48 M
29
100 M
30
697 M
31
BOT LAYER
API LAYER
SERVICES LAYER
APPLICATION LAYER
DATA LAYER
BUILT
| CUSTOMER ENGAGEMENT

BOTS CAN WORK
WELL AROSS MESSAGING
PLATFORMS
• Within chat apps, a bot is essentially a layer that retrieves information for a user or
group of users
• It can be as simple as extracting information from a database
• Or there could be some logic or complex calculations involved -
• This is where we would see the application of a technology that rolls up to the AI
classification that we just outlined such as Machine Learning or Natural Language
Processing
| CUSTOMER ENGAGEMENT

Conversational AI platform
CONVERSATIONAL AI PLATFORM
Source: MindMeld

74
PEOPLE USER  
EXPERIENCE
CHATBOT/ASSISTANT USE CASESINTERACTION
CHANNELS
Text
Voice
Graphics
Image
Video
Virtual world
GUI
Touch
Gesture
Dialogue
• Product and Service 
Information
• Product and Service
Selection and Transaction
• Trip Planning
• Arrival and Departure
• Concierge Services
• Events and Activities
• Customer Feedback
Web page
IM, Chat, SMS
E-mail
Activity stream
Smart agent
Mobile app
VR and AR
Homes
Automobiles
Wearables
IOT
Human /
Machine UX Listening (NLP)
Cloud Services
Open APIs Databases Apps Devices
Chatting (NLG)
Business Logic
Knowledge ML
| CUSTOMER ENGAGEMENT: The anatomy of chat bots and virtual assistants

Structured &
unstructured
data ingest
Context, intent and sentiment
analysis
Story & conversation
management
Speech
processing
Natural language understanding
Semantic search
Natural language
generation
Descriptive, predictive &
prescriptive analytics
Speech generation
Symbolic reasoning &
Real-time decisioning
Question answering
Recommendation
Advice
Expert assistance
Task planning
Command execution
Semantic APIs for externally provided capabilities
(UI, data, AI engines, external systems and services)
UI, task, & service
orchestration
Visualization &
presentation
Data/service 
provisioning
Knowledge acquisition
Image
processing
Knowledge
management
Image, face & gesture
understanding
Machine learning algorithms
| CUSTOMER ENGAGEMENT: The functional building blocks for chat
bots and virtual assistants

CHANNELS
BOT
FRAMEWORKS
CONVERSATIONAL
USER INTERFACE
IINTELLIGENT
ASSISTANCE
(Service APIs)
KNOWLEDGE ENGINES
DATA & SERVICES
Native Apps PlatformsWeb Browsers Mobile GuestsLive Agents
MicrosoftGoogle IBM KORE VIV
Natural Language Processing
Speech Recognition NLU: Words, Syntax, Context, Semantics, Sentiment, Personality, Intent Conversation & Dialog NLG Personality & TTS
Vision Processing
Image Recognition
Knowledge Base Semantic Engines Analytics Engines Process/Workflow Engine
Pre-Trip Exploring
Immersive Play
Discovery
Planning
Booking
Learning
Way Finding
Personalized Services
Arrival Guidance
Real-time Suggestions
Relevant Notifications
Character Concierge
Contextual Commerce
Virtual Purchases
Shopping
Sharing
Reminisce Moments
Cross-Device Messaging
User, Task and Service OrchestrationLanguage and Dialog Models, Domain
Ontology, Predictive Models, and Task Expertise
Knowledge Representation, Common Sense &
Causal Reasoning
Machine Learning and Deep Learning:
Diagnostic, Predictive, and Prescriptive
Analytics
Data Services
1st, 2nd & 3rd Party Data, Social Networks, Reference Data, RDBMS, Graph DB, CMS Marketing, CRM, ERP, MDM Systems Administration, Monitoring Dashboards
Enterprise Services and Operations
Apple Nuance
SEM
PRE
YUBIIHOUNDFacebookAPI.ai
AI: A CROWDED MARKETPLACE

Native AI platforms for chatbots
Leading internet technology companies providing native  
AI-based personal assistants and bot frameworks for building  
and deploying conversational interfaces:
77
▪ Amazon—Alexa provides voice interaction with devices and
services. Alexa Skills Kit provides a collection of self-service APIs,
tools, documentation and code samples for adding skills to Alexa.
▪ Apple—Siri artificial intelligence and natural language processing
enable conversational interface, personal context awareness and
service delegation. SiriKit lets developers integrate services with
Siri.
▪ Facebook—Facebook Bot Engine is based on WIT.ai, which trains
bots using sample conversations. API calls extract meaning and
intent from sentences.
▪ Microsoft—Cortana voice or text activated intelligent personal
assistant platform supports multiple devices, languages, and
operating environments. Microsoft Bot Framework provides APIs
and functionality needed to build, connect, manage, and publish
intelligent bots that interact conversationally.
▪ Google—Google Assistant (AI) extends Allo, Now, Hangouts, Home,
and other products into conversational 2-way dialog that
understands the users world and helps get things done. Google
provides many best of breed APIs needed for conversational UI.
▪ Samsung Viv—Virtual assistant framework with dynamic
programming to handle complex queries.

▪ Facebook M—is an instant messaging service that provides text/voice
communication and web chat.
▪ Google Hangouts— is a communications platform that includes instant
messaging, video chat, SMS, and VOIP features.
▪ Kik— is an instant messenger application (app) for mobile devices.
▪ Line—is a Japanese messaging app.
▪ Skype— is a communications platform for text, voice, and video.
▪ Slack—is a multi-environment, cloud-based platform for team and community
collaboration
▪ Telegram—is a cloud-based encrypted service for sending messages and
exchange photos, videos, stickers and files of any type.
▪ Twilio— provides infrastructure and software as a service for business
communications, enabling phones, VoIP, and messaging to be embedded into
web, desktop, and mobile software.
▪ Twitter— is a free social networking microblogging service.
▪ WeChat— is a mobile text and voice messaging communication service that
provides text messaging, hold-to-talk voice messaging, broadcast (one-to-many)
messaging, video conferencing, video games, sharing of photographs and videos,
and location sharing.
▪ WhatsApp— is a cross-platform mobile messaging app that allows exchanging
messages without SMS charges.
Messaging chatbot channels
Messaging platforms providing chat bot frameworks, APIs  
and SDKs that support building, publishing, and managing of 
chat bots and personal assistants:
78

▪ Arria— NLG Platform generates natural language by extracting
information from complex data.
▪ Narrative Science—Quill is a AI platform for NLG that converts
data to relevant information to professional prose
▪ X.ai—Amy is an AI that arranges meetings
▪ Clara Labs—Clara is an AI who schedules meetings.
▪ Conversica—is an AI platform that acts as a sales assistant to qualify
and communicate with leads.
▪ Creative Virtual— is a virtual agent platform for self-service and
hybrid AI customer support solutions. It trains by reading
manuals and other documentation.
▪ Equals3Media— is a cognitive platform for audience research,
segmentation, and media planning.
▪ Kasisto— KAI is a conversational AI platform powering virtual
assistants and smart bots across mobile, messaging, and
wearables. KAI Banking is pre-loaded with thousands of banking
intents and millions of banking sentences.
▪ Kensho— Warren is a Siri-, Watson-style intelligent investor with
significant financial services domain expertise.
▪ Ross Intelligence—ROSS is an artificially intelligent attorney that
helps power through legal research.
Enterprise chatbots and virtual assistants
3rd-parties providing and deploying enterprise chat bots and
virtual assistants that combine domain expertise, causal
reasoning, and machine learning to handle complex tasks:

3 Leading internet technology companies providing enabling technology for AI platforms and frameworks to build and
deploy virtual assistants that augment employee productivity, automate complex tasks, and improve customer
experience:
▪ API.ai—is an AI platform for bots, that provides speech-to-text, NLU,
intent recognition, context and conversation management, and
fulfillment of user requests.
▪ Artificial Solutions—Teneo is an AI platform for building enterprise-
class assistants that let people talk to apps in free-format, natural
language using speech, text, touch, or gesture.
▪ CyCorp / Lucid—is a causal reasoning platform that combines a large
common-sense ontology and knowledgeable with natural language
interfaces.
▪ IBM—Watson is a cognitive platform that enables software, services,
and apps that think, improve by learning, and discover answers and
insights to complex questions from massive amounts of data.
▪ Inbenta—is an AI platform for delivering intelligent chatbots for
customer service.
▪ IPsoft—Amelia is a cognitive agent (or digital employee) who can
take on a wide variety of service desk roles and communicate with
customers using natural language. She speaks 20 languages and
trains by reading manuals and other human readable materials.
▪ Kore—provides an enterprise-grade platform-as-a-service to build
and deploy AI-based enterprise bots on a large scale for varied
business use cases.
▪ Luminoso—is an AI platform for NLU that uses machine learning and
semantic knowledge graphs to put text into context, map concepts,
analyze sentiments, and derive insights from varied sources.
▪ Nuance— Nina is an intelligent cross-channel virtual agent platform
that converses via voice or text, and delivers instant, accurate,
successful outcomes in a natural, human-like way.
▪ Robin Labs— is an open, expandable AI platform for building
conversational virtual assistants that communicate through natural
language including speech, text, gesture, and visual imagery, learn
from examples, can accomplish tasks for a user.
80
Enabling technologies for chatbots & assistants
Leading technology companies providing enabling NLP, AI technologies
and bot frameworks to build and deploy consumer and enterprise
virtual assistants that improve customer experience, augment
employee productivity, and automate complex tasks:

PUBLICIS.SAPIENT  
AI PLATFORM FOR CUSTOMER
ENGAGEMENT
(YUBII + KAAS)
Source: Publicis•SapientThis content included for educational purposes.

YUBII + KAAS
Ubiquitous cognitive experience framework + Knowledge as a Service approach allows you to leverage best-of-
breed AI capabilities from the widest marketplace
YUBII Framework
•
• The Sapient Cognitive Experience Framework which enables
multi-channel user experiences with seamless cross channel
integration, and is built to support dynamic user experiences
through chat, image, video, AR/VR, and a flexible approach to
support new UI paradigms as they are born.
Supports Interactive Learning with the ability to monitor and
learn interactions as part of a user experience automatically
in order to minimize training and pre-configuration.
• Enables the seamless integration of live human agents and
supporting AI capabilities to deliver the greatest experience
possible.
KaaSApproach
•
•
•
•
The Sapient Knowledge as a Service approach provides an
architectural plan which enables flexible delivery of knowledge
services across a diverse set of knowledge engines, enterprise
integrations, and data sources.
Provides a single, central resource for user experiences to
establish their ability to access, modify, and interact with the
world outside of the user experience.
Provides future-proofing through it’s ability to allow migration of
integrations and knowledge engines without modification to the
user experience.
Provides a common deployment for choosing best-of-breed
solutions from the market.

| CUSTOMER ENGAGEMENT: “YUBII” accelerator for ubiquitous brand engagement

• Yubii is a Cognitive Experience Framework that orchestrates the
technologies and information needed to have an intelligent
conversation (NLP, Knowledge Engine, Experience Framework, State
Management, Live Agent, Data)
• Yubii allows us to build user experiences and deploy the underlying
knowledge models to multiple endpoints (mobile app, website,
messenger, connected device, virtual / mixed reality). 
• Yubii also allows us to manage a conversation across endpoints,
without aggravating the user by asking them to restate their
intentions. 
• Yubii is technology agnostic. It can work with multiple ML and cloud
vendors, nlp components, knowledge engines, content management
systems and analytics solutions.
YUBII: OVERVIEW
84

YUBII: VIRTUAL ASSISTANT TRAINING LIFECYCLE
Input
Collection
Data Prep &
Analysis
Conversation
Design & Training
Testing Go Live
Ongoing
Improvement
• Geo location
data
• POI details data
• Guest data
• Park operations
data
• Audio: Speech to
text processing
• Data cleaning &
normalization
• Pattern analysis
• Conversation
structure design
• Training data
creation
• Testing plan and
data creation
• Dialog
• Intents
• Entities
• Contexts
• Fulfillments
• Training
synonyms
• Fulfillment micro-
services
• Review
• Curate
• Improve
• Release new
Virtual Assistant
• Monitor
• Curate
• Improve
FEEDBACK FEEDBACK

KAAS: HARD AI / SOFT AI
Intelligence: The ability to acquire and apply knowledge and skills
Perceive Understand
Intelligence
Act Decide
Observe
Direct
The World
“SOFT AI” – Cognitive Computing
Conversation Communication
Representation Perception
Generation Production
“HARD AI” – Machine / Deep Learning
Calculation Computation
Classification Regression
Reason Memorization
Through it’s bringing together of data, enterprise integrations, knowledge engines, and user experience
frameworks, KAAS becomes an AI platform. As a system, it provides the ability to acquire and apply knowledge
and skills through the natural pairing of cognitive computing and machine / deep learning.
ACQUISITION
APPLICATION
86

COGNITIVE MARKETING, SALES &
SERVICE

ASK
88
Cognitive marketing, sales, and service:
• Customer journey focus rather than product
• Customer & prospect big data and predictive
analytics for better decisions sooner & at scale.
• Adaptive, context-aware websites and microsites
• AI-first mobile apps and seamless multichannel
user experience (smartphone, tablet, wearables)
• 1st/3rd-party message apps
• Dynamic pricing, delivery, service, and support
• Concierge QA, advice, and personalized service
• Customer digital engagement, feedback, reviews,
ratings, and loyalty programs

Rich landscape for
customer-centric
digital experience
today

Use of cognitive and analytics to drive customer engagement
Source: IBM 90

CLIENT
EXPERIENCE
Research
Options
Life
Event
Select &
Engage
Associate
F2F Digital
Enablement
Client
Portal &
Reporting
Peer
Networking /
Community
Develop
Network Customer  
Goals, Needs, 
Constraints
Custom  
Product/Service
Construction
Guidance &
Resources
Credit,
Financing & 
Payment Plan 
Enrollment &
Onboarding
Product/
Service Training
& 
Operations
Regulatory,
Risk &
Compliance
Visualization
& Reporting
Reassess
Goals, Life
Events
Modeling &
Adjustment to
Strategy &
Plan
Extended
Networks &
Resources
Networking
Support
Community
Building
Broadening the
Relationship
Optimization
& Tuning
Execution and Management
Acquire &
Understand
Client
Initial
Engagement,
Strategy &
Planning
TEAM MEMBER EXPERIENCE
Cognitively enable the
customer experience of
both clients and
employees
91

• Big data and advanced analytics give firms a host of ways to target and
connect with the kinds of investors they want.
• Marketing is shifting from supporting sales and distribution to actively
targeting clients and engaging investors in the right place at the right time
to build and nurture relationships.
• Big data and analytics are key to better identifying which marketing
strategy is best suited for a given prospect at a given time and to providing
a custom experience tailored to every individual client.
• Semantic data mining and machine learning automate gathering prospect
information such as contact information, demographics, income, interests,
and in-market behavioral signals (e.g., site navigation, emails, voice
conversations, content downloads, etc.) that prioritize leads and better
target messaging, offers and market interactions.
• Sales agents and managers can mine this trove of information to create
SEO-optimized content, customize it for specific user contexts, and tailor it
for delivery across different media such as email, social media message,
tweets, and other information channels and devices.
Precision marketing
technology gives firms a
host of ways to target and
connect with the kinds of
customers they want.

Source:AccorHotels
Source: Third Door Media
Search engine
optimization success
factors
93

• Before the rise of social platforms and interactive digital media, corporate
communication was generally a one-way street; even websites and email
are mainly broadcast media that offer limited interactivity.
• Social media exploded those limits, empowering even big, anonymous
corporations to have meaningful conversations with their customers,
employees, partners, colleagues, and the world at large.
• Even regulators have shown enthusiasm for social media, officially
recognizing their value in helping educate investors and prevent fraud.
• More advanced companies use social media to “listen at scale” to learn
customer characteristics, and to understand emotional motivators of
investor behavior and feelings, including some factors of which customers
may not be aware.
• Some firms sift tweets, emails, and voice communications from traders,
investors, and analysts for market insights and investment signals to better
inform decision-making.
Meaningful social media
conversation creates value

• AI technology helps enterprises engage the right clients with the right
offerings at the right time and through the right channels.
• AI CRM tools do more that present internal information in an organized
way. The best portfolio managers are also the best relationship managers.
• AI-based CRM tools can continuously monitor customers’ social media
posts, tweets, credit factors, and other data points and can alert
investment managers accordingly.
• AI applications can initiate event-driven personalized communications
with customers, and engage them in near-human ways that traditional
software cannot.
• AI-based CRM platforms can interface with customer portals to provide
customized user interfaces. The customer is presented with the
information he or she is most likely to need, based not only on previous
interactions, but also on big-data predictive analysis.
AI CRM service
personalization

• Enterprises need infrastructure that provides instant access to account
data and documents, a 360-degree view of their assets, rapid information
processing, and effective tools for easy access to self-service research and
advice.
• AI-enhanced content is provided through a knowledge base or resource
center that clients can access at any time to get the insights and latest
research they need to either inform their own decisions or drive
discussions with their human and digital agents.
• The data that results from clients and prospective clients accessing and
downloading specific content assets can provide deeper insights into the
specific needs of each individual, enabling managers to reach out with
precise messaging that answers their most pertinent questions – without
clients ever having to ask. It’s this level of personalized service that
enables enterprises to lead the competition in the digital age.
• As speech processing and natural language processing technologies
mature, AI applications handle many customer service queries without
human involvement.
AI applications handle
many customer service
inquiries using speech
processing and natural
language processing

Marketing technology landscape 2016
97

Precision marketing company briefs and case examples*
• Company briefs and case examples highlight vendors that
provide predicbve markebng solubons for B2B and B2C
customers.
• Vendors enable access to aggregated data sets of enbbes,
individuals, and behavior from internet and other sources.
• All provide some level of pre-packaged predicbve models
and DIY training.
• All develop predicbve models through machine learning
that analyzes 1st, 2nd & 3rd party data and historical
conversion outcomes to some extent.
• Most provide demand generabon and lead scoring based on
predicbve models as an alternabve to hand-built models.
• Martech players provide interfaces to markebng automabon
and CRM packages — e.g., HubSpot, Marketo, Salesforce.
• More compebtors and potenbal partners exist if we
generalize the POC pla…orm concepts to handle more types
of life events and more industry segments.
• The slide depicbng vendors providing machine intelligence
for marke+ng idenbﬁes 150 vendors in 24 categories.
• In parbcular, we may see compebbon from vendors of
customer analybcs and advanced analybcs as covered in
Forrester and Gartner reports, and summarized in two
charts.
• The highlighted vendors provide toolsets, workbenches,
pla…orms, integrated environments and whole solubons
that can support precision markebng.
• Capabilibes provided vary, but can include descripbve
analybcs, predicbve modeling, prescripbve analybcs, data
mining, text analybcs, forecasbng, opbmizabon, simulabon.
98
* Not part of this research deck

24/7
6Sense
Adobe
AgilOne
Aginity
Alteryx
Angoss
AYASDI
Bluecore
BlueShift
BlueYonder
Datacratic
DataMentors
Deloitte
DynamicYield
Emarsys
EverString
FICO
Grey Jean
IBM
Infer
KNIME
Lattice Engines
LeadSpace
Mintigo
Oracle
Pitney Bowes
RapidMiner
Radius
Reach Analytics
SalesForce
SalesPredict
SAP
SAS
Teradata
Tiny clues
Versium
WealthEngine
wise-io
99
Precision marketing*
• Machine learning
• Descriptive analytics
• Predictive analytics
• Prescriptive analytics

BUSINESS ACCELERATION
+ OPTIMIZATION

• Business acceleration
• Cognitive enterprise
• Cognitive platform
• Intelligent automation
• Intelligent ecosystemsOverview of 
Business acceleration  
+ optimization

busi·ness ac·cel·er·a·tion + op·ti·mi·za·tion
/ˈbiznəs//akˌseləˈrāSH(ə)n/+/ˌäptəməˈzāSHən,ˌäptəˌmīˈzāSHən/
Business acceleration is the automation of knowledge generation that
drives cost savings, competitive advantage, and new business lines
through smarter deployment of resources.
Optimization is the action of making the best or most effective use of a
situation or resource.
For example, using machines to replicate human actions and judgment
with robotics and cognitive technologies, automating repeatable tasks to
improve efficiency, quality, and accuracy of processes, while lowering
costs and freeing profits and revenue from the scale constraints of manual
labor.
What is  
business acceleration?

cog·ni·tive en·ter·prise
/ˈkäɡnədiv//ˈen(t)ərˌprīz/
Enterprise refers to a project or undertaking,
typically that is difficult or requires effort.
Cognitive enterprise is the future of public
and private sector businesses or organizations
that utilize knowledge acquired by artificial
intelligence and digital technologies to better
understand and respond to customer,
colleague, citizen, and/or stakeholder needs;
provide products, services and information
digitally; and improve operations to reduce
cost, drive revenue, and maintain compliance.

Cognitive enterprise is based on machine learning, natural language
processing, and intelligent human interface technologies.
• An enterprise’s cognitive systems can learn and build knowledge from various
structured and unstructured sources information. They can understand
natural language and can easily interact with users, other devices, and other
data sources.
• To illustrate, instead of surfing web pages, simply talk into a single input box,
e.g.: “I lost my card.” A quick chat with a rep (you didn’t even notice was not
human) and a new card is on its way. A cognitive enterprise provides
consistent and personalized service.
• Further, cognitive enterprises leverage machine learning and big data to
predict customer needs (e.g., it interprets based on analyzing a lifetime of
customer data, web data, and social media) and proactively suggest a
personalized product or service. It can do this at a scale not possible with
manual only methods, and it can learn and improve as it handles more cases.
• Cognitive systems capture the expertise of top performers, accelerate
development of expertise in others, and enhance the decision-making of
professionals across the enterprise.
Cognitive enterprise is
based on machine
learning, natural
language processing,
and intelligent human
interface technologies
$

9
Acquired Sqream, which uses
machine learning to detect
behaviour patterns of wealth
customers
Goldman Sachs invested $15m to help fund
Kensho, the natural language search engine
designed to analyse news events and answer
detailed questions about financialmarkets
ING mobile app allows
transactions to be made
through voiceactivation
Sage has developed a chatbot
called Pegg that acts as a
business accounting personal
assistant
Zest Finance – AI
underwriting which offers
40% improvement over best
in class industryscore
Siftscience – can helptheir
clients detect 89% of fraud
while reviewing only 1% of
customercases
Digit – an automated savings app
that reviews your spending habits
and proactively saves money you
canafford
106
COGNITIVE ENTERPRISE IS ALREADY HERE

what’s my Quicksilver
card balance?”
EXAMPLE:
ENGAGING
CUSTOMERS
Capital One has deployed a new skill to Amazon Alexa that
powers voice activated banking. This application is already
integrated with transactional systemspermitting payments in
addition to balancequeries.
Manage your Capital One accounts using nothing but your voice
Credit Cards
• Checkbalance
• Get duedates
• Pay Capital One cardbill
Checking and Savings
• Checkbalances
• Review recenttransactions
Auto Finance (New) Home Loans (New)
• Check principal balance • Check principal balance
• Get payoff quote • Get due dates
• Make a Capital One payment • Make a Capital One payment
“Alexa, ask Capital One

IS BECOMING THE NORM
Dominos Pizza
one of many fooddelivery
chatbots
CognitiveCommerce CognitiveService
Donotpay
the world’s first robot lawyer
that hasoverturned 160k
parkingfines
Uber
cognitive commerce via chat
integration with Googlemaps
Twyla
AI driven support chatbot that learnsfrom
human agents in order to improve FAQ content
which is only 50% effective
H&M
personal stylist chatbot, creating a
service that was uneconomical with
humans
Ivy
Go Moment’s hospitality systemcapable of
handling 90% of guestrequests
12
COGNITIVE BUSINESS
108

Artificial Intelligence + Experience Design
Logic Magic
Only by addressing both elements can compelling experiences be ones that surprise and
delight customers and colleagues, making the bank feel more human.
109
COGNITIVE ENTERPRISE SITS AT THE
INTERSECTION OF LOGIC AND MAGIC

cog·ni·tive plat·form
/ˈkäɡnədiv//ˈplatfôrm/
A cognitive platform is a software platform based on the scientific
disciplines of artificial intelligence and signal processing that encompass
machine learning, reasoning, natural language processing, speech and
vision, human-computer interaction, dialog and narrative generation,
and more.
A software platform is a major piece of software, such as an operating
system, an operating environment, or a database, under which various
smaller application programs can be designed to run, together with
externally provided devices and services that are provisioned using
application program interfaces (APIs).
What is a  
cognitive platform?

REASONING
ENGINES
INTEGRATION &
CORE SERVICES
KNOWLEDGE SERVICES
SERVICES API (KNOWLDEGE ENGINE ABSTRACTION LAYER)
Free-Text Search
Contextual
Search
Graph
Search
Reasoning /  
Explanation Logging
Caching
Security
Monitoring
KNOWLEDGE BASE MGMT
Ontology
Creation
Ontology
Evolution
Concepts/Entities
Extraction
Relations/Fact
Extraction
CONTENT DELIVERY
Query Parser Inference
Deductive
Relevance-based
Results Ranking
CONTENT INGESTION
Indexing
Content Storage
Tagging/Metadata
Extraction
Facets/Filters
Concepts
Disambiguation
Results post-
processor
User Entitlements
Check
Query
Formulation/
Expansion
Structured 
Data Access
Text Generation
User
Entitlements
Indexes,
Structured Data
Triples
Knowledge
Base, Reasoning
Topic Modelling, Query
Expansion, NLP/NLU,
Ontology
USER EXPERIENCE FRAMEWORK
TEXT, AUDIO, VIDEO, BIOMETRIC, AND IOT SENSOR INTERFACES
ASR / TTS ServiceMobile framework
Text Interfaces
Voice & Tone
REST / SOAP
APIs
Managed File
Transfer
Ontology
Natural
Language
Translation
NLP / NLU
Audio/Video Capture Audio/Visual Output
Reporting / Portal Email
Heuristic
MACHINE LEARNING
Deep
Learning
Engines
Neural
Networks
Constrained
Conditional
Models
Deep
Learning
Engines
Chat Agents
NLG
This diagram depicts
functionality to
support cognitive
enterprise. A business
would incorporate
various portions of
this architecture in
phase, for example  
a crawl, walk, run
approach.
SaaS
AI Layer
PaaS
IaaS
PUBLIC/HYBRID  
CLOUD STRUCTURE
Example business context diagram for a cognitive enterprise
112
This content included for 
educational purposes.

Cognitive enterprise business context diagram — partner overlays
This overlay to the cognitive enterprise business context diagram depicts where  
partner platforms might be deployed to address various functions.
Indexing Tagging/Metadata Text Generation Facets/Filters
Extraction
Structured Data
Query Parser
Relevance-based
Access Content Storage ResultsRanking
Results UserEntitlements
Post-Processor Check
CONTENT INGESTION CONTENT DELIVERY
Ontology Creation
Concepts/Entities
Extraction
Ontology Evolution Relations/Fact
Extraction
KNOWLEDGE BASE MGMT
Heuristic
Inference
Deductive
REASONING
ENGINES
Text Interfaces Audio Capture AudioOutput
TEXT & AUDIO INTERFACES
Chat Agents Voice & Tone Mobile Framework ASR / TTS Service Reporting / Portal Email
USER EXPERIENCE FRAMEWORK
Concepts Reasoning/
Disambiguation
Free-Text Search Contextual Search Graph Search
Explanation
Query Formulation Query Expansion NLP /NLU
SERVICES API (KNOWLEDGE ENGINE ABSTRACTION LAYER)
Ontology
Indexes, Structured Knowledge Base, Topic Modelling, NaturalLanguage
Data Triples Reasoning Query Exapansion, Translation
NLP/NLU,Ontology
KNOWLEDGE SERVICES
Deep Learning Constrained
Engines Neural Networks ConditionalModels
MACHINE LEARNING
REST / SOAPAPIs
ManagedFile
Transfer
Security
Logging
Caching
Monitoring User
Entitlements
INTEGRATION &
CORE SERVICES
BESPOKE
AND POINT SOLUTIONS
SEMANTIC AI
ML
PLATFORMS
SEMANTIC AI
ALIGN WITH
INTERNAL
ARCH
ML PLATFORMS
NATURAL LANGUAGE &
VISION PLATFORMS
NATURAL LANGUAGE AND SEMANTIC PLATFORMS
113

Enterprise AI platforms company briefs and case examples*
• Company briefs and case examples highlight trends toward enterprise AI and
cognibve pla…orms in the following sectors:
- Internet & compuGng — Amazon, Apple, Baidu, Facebook, Intel, Nvidia,
Google Alphabet, Microsoj, HPE. Deep learning, proacbve agents, open
source pla…orms
- IT services (BPOs and consultancies) — TCS, Infosys, Wipro, Mphasis,
Accenture, Booz Allen, Deloihe, EY, KPMG, PwC. AI pla…orms for robobc
automabon, cognibve compubng, intelligence augmentabon for knowledge
intensive services.
- Financial services — Cib Group, Goldman Sachs, USAA, UBS, Amex,
MasterCard, BofAML. AI for ﬁnch innovabon, business disrupbon, and
collaborabve services automabon.
- Retail — Amazon, Lowes, Staples, Target, NorthFace, WayBlazer. Customer-
centric predicbve markebng, cognibve travel services, virtual assistance.
- Manufacturing — Siemens, Fujitsu, Hitachi, Maana, General Electric, Ford,
Toyota, General Motors, Tesla. AI & cognibve pla…orms for engineering,
manufacturing, and product lifecycle management
114

Accenture
Amazon
Apple
Baidu
Booz Allen
CYC/Lucid.AI
Dato
Deloitte
Enterra
EY
Facebook
Fujitsu
GE
Google
HPE
Hitachi
IBM
Infosys
IPsoft
KPMG
Maana
Microsoft
Palantir
PWC
Rage Frameworks
Siemens
Skytree
TCS
WayBlazer
Wipro
WorkFusion
Enterprise AI
platforms*
115

a good rule of
thumb for
automating
knowledge work
100:1
NEW
BUSINESS
• Intelligent search
• Broader coverage
ENHANCEDINSIGHTS
• Automation of manual activities
• Conversion of unstructured dataCOST REDUCTION
• Reduce time required
• Scale human effectiveness
COMPETITIVEADVANTAGE
| BUSINESS ACCELERATION  
+ OPTIMIZATION
!117This content included for educational purposes.

Big data to intelligent applications: a lifecycle view
118
DATA INGESTION
Data preparation
• Data integration
• Data enrichment
• Data imputation
• Data versioning
• Data provenance
• etc.
Natural language
processing
• Entity extraction
• Entity resolution
• Relationship extraction
• Taxonomy generation
BIG DATA
Web content 
(web sites, blogs, …)
Social networks 
(Twitter, Facebook,…)
Online activities 
(Search, shop, games…)
Enterprise apps 
(ERP, CRM, …)
Internet of things 
(Sensor, device data…)
Processes 
(logs, data lineage,…)
Textual content 
(Documents, reports, …)
Knowledge-bases 
(taxonomies, ontologies,…)
SEMANTIC GRAPH MACHINE REASONING
Sensemaking engine
Recommendation engine
Process automation
engine
Context engine
Semantic search
Inference engine
Rule engine
Semantic query engine
Machine learning  
(classification, clustering,
anomaly detection
INTELLIGENT APPLICATION
Find 
(people, content, …)
Compare 
(products, companies, …)
Detect 
(incident, anomaly,
opportunity, …)
Discover 
(Insight, pattern, …)
Analyze 
(Performance, problem, …)
Design 
(Product, svc, process…)
Predict 
(demand, inventory, …)
Prescribe 
(Next best action, …)
Network of:
people,
places,
organizations,
processes,
rules,
policies,
events,
documents,
devices, etc.
Semantic inferencing
Learning from usage patterns
Automated  
update cycle

BOT 1  
PRODUCT SELECTION
BOT 4  
GIFTING
BOT 3
HOW-TO CONTENT
BOT 2  
COMMERCE
BOT 5  
REGISTRY
personal assistant
(the conductor)

Intelligent collaboraGon
Intelligent processes are  
goal-oriented & event-driven.
Processes adapt and self-
opbmize when events
happen, excepbons occur, or
needs change. Emergent
projects learn. Knowledge
models evolve.
Fixed
Transacbon
Dynamic 
Case
Emergent 
Project
VALUE
KNOWLEDGE INTENSIVITYLow Hi
LowHi
• Semanbc data models, process
models, and rules connect Info
and systems across
organizabons, jurisdicbons,
and geographies.
• Goal-oriented acbvibes to perform
• Decisions required to take acbon
• Rules & condibons to be met to choose
• Data & calculabons determine condibons
• Semanbc, machine learning, model-
driven methodologies for knowledge-
and data-intensive collaborabve
knowledge work.
• Authoring, collaborabon, analysis, and
communicabon tools are semanbc.
• Design = Model = Applicabon =
Explanabon = Documentabon.  
(Model executes directly)
• System learns, project models evolve
• Simulabon tesbng & automated version
control are nabve.

The fastest workﬂow travels the fewest steps,
touches the fewest hands, and does as much  
as possible for you.

PERSONALIZED AND
FRICTIONLESS
EXPERIENCES
Companies are using real time and
customer personal data to
automate and predict needs
starting from prompting the dream
phase, to awareness, engagement,
and every step of the customer
journey.
DIGITAL PLATFORM
TRANSFORMATION
Companies are replacing old or
homegrown systems with digital
platforms that enable seamless
experiences, support customer
services and drive business
decisions.
BUSINESS MODEL
EVOLUTION
Companies are expanding their
products and new services to better
the serve the needs of the new
customer expecting the the ultimate
experiences at the touch of a finger.
Three business acceleration power moves
122

BACK STAGE FRONT STAGE
Organization 
& 
Operations
Service 
Value Chain
LINE OF VISIBILITY
OmniChannel 
Communication
User 
Experience
Customer  
& Employee
Experience
3rd
Party
Partner
Experience
Distribution
Partner Experience
B2B and B2B2C
Enterprise services  
cognitive transformation  
and business acceleration  
framework

Cognitive Business

Recomendados

Recomendados

Mais conteúdo relacionado

Semelhante a Cognitive Business

Semelhante a Cognitive Business (20)

Mais de Mills Davis

Mais de Mills Davis (6)

Último

Último (20)

Cognitive Business