Nasscom Big Data and Analytics Summit 2016: Session III: How Cognitive Computing can Rev Up your Enterprise - A CIO Perspective
1. Sri Donthi
SVP and CIO for AMENA, PepsiCo Inc.
23rd June 2016
All copyrights and trademarks referenced herein are the property of their respective holders. The opinions expressed in this presentation are those of the author and
do not reflect those of PepsiCo or any referenced third parties
Promise of Cognitive Computing in
Consumer Centric World
2. Imagine…..If machines could be
taught to leverage data, learn
from it, and, with a little guidance,
figure out what to do with it?
4. Cognitive systems will transform industries by providing
powerful analytics and insight to the end user
5. VOLATILITY
The dynamics and
nature of change . . .
and the rate / speed
of that change
Lack of (and likelihood)
of surprises
predictability, the
possibility
UNCERTAINTY COMPLEXITY
Multiplicity of problems
and issues, some
uncontrollable, causing
chaos and confusion
AMBIGUITY
Mixed messages,
intangible/unclear goals,
poor communication,
lack of meaning
World we live in…
6. Data explosion in connected world…
2.5 quintillion bytes of
data created every day
90% of the data in the
world today has been
created in the last two
years alone
Every minute 1.7
megabytes of data is
created for every person
on the planet. All 7.3Bn
of us.
….and imagine the possibilities
Retail
Consumers post 500 million tweets and
55 million Facebook updates each day
Source: IBM
7. Ever changing consumer…
Increasing value conscious but
tech-savvy and demanding
customers
Growth in the use of consumers
using digital and mobile devices
to shop
(reading reviews / comparing prices)
Confluence of these trends is the signal to use digital intelligence to drive superior growth
Consumer buying behaviour can
be very complex
9. Hyper Performance 1:1
Consumer Engagement
81%
of companies say they
have or are close to
having a holistic view
of their customers
37%
of consumers say
their favourite retailer
understands them
Personalized shopping experience
Discover unseen pattern
Be responsive to consumer needs
Efficient merchandizing
Efficient supply chain
Bring differentiation
10. Differentiating consumer experience
using digital intelligence
Better collaboration
enabling seamless
experience across
channels
Decide
Digest vast amounts
of data to identify
and implement new
ideas
Evidence-backed
recommendations,
with BMC, cost
structures and
customer behavior
Discover
Enhancing the speed
of processes, quality
and integrity of
products
AutomateEngage
Improving
internal processes
What does it mean for CPG Companies
Source: IBM
11. Agricultural Intelligence
Multi-sensor agriculture systems enable an ethical, efficient & high yield process
• 28% global population directly or
indirectly employed in agriculture
• Worldwide agriculture accounts for 70%
of all water consumption
• Weather disasters have caused many
failed harvests in the past decade
• Greatest challenge our civilization has
ever faced is feeding 8bn people in a
sustainable way
13. Cyber Security
Building instincts and expertise, supporting analysts outthink and outpace threats
• 80% security information unstructured and
inaccessible for traditional systems
• 75,000+ documented software vulnerabilities
• 10,000+ security research papers issued each year
• 60,000+ blogs published a month
Source: ISACA Source: IBM
14. Creating valuable consumer experience
Enhancing Products
and Services
Automating
Processes
Uncovering Insights
Ease of use
and
convenience
Simplicity
Emotional
effects
Empower
consumers and
instill
purchasing
confidence
Benefits
15. Imagine the future…..
Gather information
about consumer
preferences
Engage consumers at
each touch-point of the
brand experience
Offer products to meet
consumers preference
Provide consistent
brand experience
in all channels
16. Our approach…
Develop a roadmap
(Dream BIG)
Experiment small
(Fail/Learn Fast and Cheap)
Deploy, explore and evolve
(Aggressively Deploy)
Develop and train the models
(Key Opportunities)
Source: IBM
Editor's Notes
Image: Shutterstock
Notes:
Cognitive system can be self-learning. They are like bright students who are given educational material and they can learn by themselves
Image Source: Shutterstock
Cognitive computing technologies and solutions also face several obstacles. As with any new technology promising to change the world, business leaders wonder how to separate hype from real potential. But, executives shouldn’t let these concern obscure considerable likely benefits of cognitive computing.
Examples of how cognitive technologies can be integrated into compelling business solutions.
Computer vision: The ability of computers to identify objects, scenes, and activities in unconstrained (i.e naturalistic) visual environments.
Natural Language processing: The ability of computers to work with the text the way humans do. For instance, extracting meaning from text or even generating text that is readable, stylistically natural and grammatically correct.
Machine Learning: The ability of computer system to improve their performance by exposure to data without the need to follow explicitly programmed instructions.
Image Source: Shutterstock
Image Source Shoppers: Shutterstock
Big Data Image and Facts: IBM
Notes:
Digital technologies are reshaping both consumer demand and competitive dynamics in the CPG.
CPG is fast appraoching a tipping point. From now on, technology enabled innovations will drive much of the growth in CPG.
Majority of data produced today is termed ‘Unstructured Data, this accounts for 80% of all data generated today and is only expected to grow to over 93% by 2020.
Every minute, 100 hours of video are uploaded to YouTube, and more than 100 Billion Google searches are done every month.
But what do we do with all of this data?
Consumer Engagement : Image of person: Shutterstock, Diagram: IBM
Agricultural Intelligence: blogs.worldbank.org
Cyber Security Image : Shutterstock
Smart Sales Execution: Shutterstock
Notes:
Every customer is unique and has little tolerance for businesses that fail to recognize their specific interests, wants and needs, Cognitive systems will help transform how companies serve each customers. Busineesses will be able to understand, reason and laearn from every customer interaction.
Image Source Person: Shutterstock
Notes:
The days are over when retailers can target promotion to mass audiences based upon age, gender and income. Customer relationships are no longer built on isolated interactions. Its an on-going dialoge that must span touch points in the customers journey.
There is a big divide between what companies think they know about their consumers and what consumers think the companies know about them. Hyper connected consumers expect a lot more from brands and are ready to share their information with them only if they could get personalized advice, recommendations from the companies. CPG companies must therefore look to reduce this gap.
Anticipated benefits of 1:1 targeting
Better consumer engagement leads to better conversion
Discover unseen pattern hidden in their data
Up-sell opportunities
Be more responsive to your customers with fast and personalized sericves across all channels
Bring data sources with merchadising to drive sales, margins and differentiation
Stay ahead of competition with innovative ideas and future forward technologies that enhahnce customer experience and provide hyper personalization, ultra convinience, and a simplified shopping experience
The retail industry is experiencing unprecedented change. Over the past decade, emergence of technology-enabled “smarter consumers” has upended traditional retail business models and changed consumers expectations.
As the mountain of data continues to grow, existing analytics capabilities are not sufficient to gain the necessary insights to fully meet ever-growing and ever-changing consumer desires.
Gone are the days when retailers can target promotions to mass . Today, there is only one channel that counts: the consumer. And consumers expect retailers to be relevant and responsive.
The impact of using data can be substantial translating into bottom line.
Retailers who understand this and redefine their data management strategy will stay ahead of
competition in a world that is increasingly becoming data dependent.
Understand, connect and engage customers by delivering personalized shopping experiences
Build efficient merchandising and supply networks to meet the demands of empowered customers
Transform and optimize operations to improve performance and operational efficiency
Increase customers' spend and store visits by building customer- focused initiatives and marketing techniques
Be more responsive to your customers with fast and personalized services across all channels.
Stay ahead of competition with innovative ideas and future-forward technologies that enhance customer experience and provide hyper-personalization, ultra convenience, and a simplified shopping experience
Bring together data sources with merchandising to drive sales, margins and differentiation
Image Source: Shutterstock
We see three broad areas of capability for cognitive systems. Opening new doors for innovations, these capability areas directly relate to the ways people think and work and demonstrate increasing levels of cognitive capability. In the future, we will see systems with higher-orders of cognitive capability. It is important to note that these capabilities are not mutually exclusive. A specific business solution may in fact leverage one or more of these capability areas.
Engagement – These systems fundamentally change the way humans and systems interact and significantly extend the capabilities of humans by leveraging their ability to provide expert assistance and to understand. These systems provide expert assistance by developing deep domain insight and bringing this information to people in a timely, natural and usable way. Here, cognitive systems play the role of an assistant – albeit one who is tireless, can consume vast amounts of structured and unstructured information, can reconcile ambiguous and even self-contradictory data, and can learn. In this partnership, the two – human and machine – are more effective than either one alone. Much like the human brain, these systems begin to build models of themselves and the world around them. This world consists of the system itself, the knowledge ingested from information corpora and the users of the system. The models include the contextual relationships between various entities in a system’s world that enable it to form hypotheses and arguments. As a result, these systems are able to engage in deep dialogue with humans. Significant and proven capabilities have been built around this capability area. In the future, it is expected that increasingly more domain-specific question and answer (Q&A) systems will emerge. Many of them are likely to be pre-trained with domain knowledge for quick adoption in different business-specific applications. Additionally, future cognitive systems will advance to have free form dialogue and reasoning capabilities. (See USAA case study.)
Decision – These systems have decision-making capabilities, to the degree humans can trust and rely on their judgment. Decisions made by cognitive systems are bias-free; however, certain standards are required for humans to fully trust their decisions. Currently, cognitive computing systems perform more as advisors by suggesting a set of options to human users, who ultimately make the final decisions. (See WellPoint Case Study.) Confidence in a cognitive system’s ability to make decisions autonomously without humans will depend on the ability to query and have traceability to audit why a particular decision was made, as well as improved confidence scores of a system’s responses. A confidence score is the quantitative value produced by a system representing the merit of a decision after evaluating multiple options.
Discovery – Discovery is the epitome of cognitive capability. These systems can discover insights that perhaps could not be discovered by even the most brilliant human beings. Discovery involves finding insights and connections and understanding the vast amounts of information available around the world. With increasingly more volumes of data, there is a clear need for systems that help exploit information more effectively than humans could on their own. While still relatively immature, some early discovery capabilities have emerged, and the value propositions for future applications are compelling. Early advances in this capability area have been made in specific domains, such as medical research, where robust corpora of information exist. (See Baylor College of Medicine Case Study.)
Image source: blogs.worldbank.org
World Resource Institute – 28% Global population – source http://ow.ly/rpfMN
New York Times – Weather Comment
Jonathan Foley University of Minnesota – 8bn People
Cognitive computing is
Minimizing use of pesticides
Reducing water and other resource usage
Reducing waster
Improving quality of end products
Understanding crop behavior through analysis of images pictures and video taken.
Enterprise have plenty of sales and performance data and have several causal factors that impact the performance. For large multi national companies, it takes time to answer questions like “why”?
It can be answered today, but can be time consuming. Many times the factors impacting the performance are finite in nature, but despite that, the analysts have to go through a repeatable process to identify potential causes to explain the performance in order to be accurate.
What if this repeatability is provided to machines and through machine learning, they could bubble up the patterns and the factors causing the performance issue almost instantaneously. What if these insights are provided in an easily consumable natural language format? This will evolve to a smarter sales execution through machine learned recommendations and growth in sales revenue.
Increased Efficiency in sales execution cycle
Better and accurate decision making
Unified Insights
Recommendation communicated to the end user though one of the cognitive services. e.g. Text to Speech (or vice versa)
ELABORATE ON TECHNOLOGIES
LIKE NARRATIVE SCIENCE, VISUALIZATION AND RULE-BASED ANALYTICS
Increased Efficiency in sales execution cycle
Better and accurate decision making
Unified Insights
Recommendation communicated to the end user though one of the cognitive services. e.g. Text to Speech (or vice versa)
Cognitive computing is
Minimizing use of pesticides
Reducing water and other resource usage
Reducing waster
Improving quality of end products
Understanding crop behavior through analysis of images pictures and video taken.
Image Source: Shutterstock
Source: Deloitte University Press – The thinker and shopper
Simplicity
Fewer steps, Hands-free
No arcane demands
Real-time, natural conversations
Enables self-service
Anywhere, 24/7 usability
Ease of Use and convenience
Fewer knobs, dials, settings
Seamless, pleasing design
Reduces complexity
Learns, adapts
Anticipates/proactively meet needs
Empower consumers & instil purchasing confidence
Personalizes shopping experience
Informs, recommends
Encourages satisfaction with choices
Reduces returns and their costs
Emotional effects
Amuses, entertains
Engages
Provides companionship
Promotes well-being
Fosters affinity and loyalty
Cognitive Technologies
Pepsi/Frito Lay Pongr image recognition technology
Aether’s Cone
L’Oreal’s Makeup Genius
Aldebaran Robot
Image Source: Shutterstock
I WOULD FOCUS ON HOW KNOWLEDGE CAN BE CAPTURED, ENCAPSULATED AND LEVERAGE
TO REDUCE SPEED TO MARKET, MAKE INNOVATION MORE RELEVANT TO CONSUMERS, IMPROVE QUALITY, IMPROVE OUT-OF-STOCKS AND CUSTOMER SERVICE
AND IMPROVE COMPETITIVE ACTIONS
Know everything they can about lifestyle of consumers because “machines” have learned their habits, preferences and behaviours
Be able to engage with consumers by means of voice, videos and written communications at every touch-point of the brand experience
Make products and package them to a consumer’s preference because cognitive computing has solved sourcing, manufacturing and logistics challenges
Provide the brand experience without retail shops because cognitive computing will enable brands to learn what products consumers want and when to deliver to them
Image Source: Shutterstock
A FEW IDEAS
DREAM BIG
FAIL/LEARN FAST AND CHEAP
FOCUSED ENERGY ON KEY OPPORTUNITIES
AGGRESSIVELY DEPLOY
REINVENT AGAIN
Anticipated benefits of 1:1 targeting
Better consumer engagement leads to better conversion
Up-sell opportunities
Win-win for consumers and brands