Project Report on Artificial Intelligence in Retail

1. i
A Project Report
On
“Artificial Intelligence in Retail”
BY
“Snehal Nemane”
Under the Guidance of
“Prof. Viraj Atre”
Submitted
In partial fulfillment of the requirement for the award of the degree
PGDM
Through
iFEEL

2. i
DECLARATION
I, Snehal Nemane hereby declare that “Artificial Intelligence in Retail” is the result of the project
work carried out by me under the guidance of Prof. Viraj Atre in partial fulfillment for the award
of Post Graduate Diploma in Management.
I also declare that this project is the outcome of my own efforts and that it has not been submitted
to any other university or Institute for the award of any other degree or Diploma or Certificate.
The material borrowed from other source and incorporated in the Project has been duly
acknowledged and/or referenced.
I understand that I myself could be held responsible and accountable for plagiarism, if any,
detected later on.
This declaration will hold good and in my wise belief with full Consciousness
Place: Lonavala Name : Snehal R Nemane
Date: 8th
April, 2019 Roll No: H-91

3. ii
CERTIFICATE
Date: 8th
April, 2019
This is to certify that the Project titled “Artificial Intelligence in Retail” is a bonafide record of
work done by Ms. Snehal Nemane in partial fulfillment for the award of the Post Graduate
Diploma in Management under my supervision. The report has not been submitted earlier either
to any University /Institution for the fulfillment of the requirement of a course of study.
SIGNATURE OF GUIDE SIGNATURE OF DEAN
(Prof. Viraj Atre) (Prof. Sudhir Salunke)
DATE: 8th
April, 2019 DATE: 8th
April, 2019

4. iii
ACKNOWLEDGEMENT
This report is a complete description of my Capstone project carried out as compulsory
component of the PGDM program at Institute for Future Education Entrepreneurship and
Leadership (iFEEL). Since I am interested in Operations Management and Information
Technology especially Artificial Intelligence, the work was concentrated on application of
Artificial Intelligence in Retail in present and future. At the beginning of this project I formulated
several learning goals, which I wanted to achieve:
 to understand the basic of Artificial Intelligence and how it works;
 to see to what extent Artificial Intelligence can be applicable in present and future ;
 to use Artificial Intelligence in Retail and see how it can help retailers;
 to see how Artificial Intelligence can improve Customer experience and help to run
business efficiently.
This project report contains my activities that have contributed to achieve a number of my stated
goals. In the following chapter a description of the Artificial Intelligence and its application in
Retail is given. Finally,
I had given a conclusion on the project experience according to my learning goals.

5. iv
ABSTRACT
Artificial Intelligence is a science of making a computer, a robot or a product to think as
smart a human think. Artificial Intelligence (AI) can be said as a study of how human brain think,
learn and work when it tries to solve the problems. It was introduced to develop and create
“machines that thinks” which are capable of learning, making decisions, mimicking and replacing
the human intelligence. The aim of AI is to improve the computer functions which related to human
intelligence e.g. learning, reasoning and problem solving. There are various application of AI in
various sectors and multiple industries. To exploit full potential of AI in Retail, this research report
explores sub-fields of AI that can be apply and use in Retail industry.
This study shows how retailers can use AI as their part of their operations and with the help of
automation technology as a part of operations and decision making process it can be used to
improve cost savings, increase productivity and revenue, faster resolution to the business problem
and identification of new stream revenue. Thus by finding how AI can be applied in retail will help
to explore the full potential of AI in retail industry.
For proving the impact of Artificial Intelligence in improving the sales and better user experience,
the company named Lenskart is been taken as example. The survey is done by asking questions
related to user experiences with Artificial Intelligence as technology to the customers. Later, the
further suggestions are given as per the research done.

6. v
TABLE OF CONTENTS
Sr. No Item Page No
1 Declaration i
2 Certificate ii
3 Acknowledgement iii
4 Abstract iv
5 Table of Contents v
6 Index vi
7 List of Tables vii
8 List of Graphs/ Diagrams vii
9 Chapters 1 1
10 Chapters 2 6
11 Chapters 3 11
12 Chapters 4 18
13 Chapters 5 20
Conclusion 39
Bibliography 40

7. vi
INDEX
Sr. No Contents Page
1 Introduction 1
1.1 Introduction to Artificial Intelligence 1
1.2 Artificial Intelligence in Retail 3
2 Literature Review 6
2.1 Different Types of AI 6
2.2 Basic Components of Artificial Intelligence 6
3 Company Profile / Industry Profile 11
3.1 Retail Industry 11
3.2 AI Industry 15
3.3 Company Profile-LENSKART 16
4. Research Methodology 18
4.1 Data Collection Methods 18
4.2 Sample Design and Techniques 19
5 Data Analysis and Hypothesis Testing 20
5.1 Data Analysis 20
5.2 Hypothesis testing 27
6 Findings and Recommendations 30
6.1 Findings 30
6.2 Recommendations 30

8. vii
LIST OF TABLES
Table No. Title Page No.
3.1 Global Retail Development Index, 2017 12
5.1 Website Engagement (Online Eyewear Retailers) 28
5.2 Revenue Information of Lenskart 28
LIST OF FIGURES/GRAPHS
Figure No Title Page No.
1.1 Artificial Intelligence different fields 2
2.1 Basic Components of Artificial Intelligence 6
2.2 Machine Learning Process 7
3.1 Smart Phone Penetration by Country,2016 12
3.2 Revenues from the artificial intelligence (AI) market
worldwide from 2016 to 2025 (in million U.S. dollars)
15
3.3 Current Adoption and Future demand of AI 16
3.4 Key Highlights of Lenskart 17
5.1 Lenskart 3D Try On Feature 20
5.2 Google Trends for Lenskart 27
5.3 Top Competitors of Lenskart 28
5.4 What AI can bring to Retail 31
5.5 Moda Polso Customers experiment 32
5.6 Image recognition technology on mobile 33

9. viii
5.7 LoweBot Autonomous Robot 34
5.8 Usage of AI in Supply Chain Management 35
5.9 GWYN Chatbot 37
5.10 Amazon Go Store 38

10. 1
CHAPTER 1
INTRODUCTION
1.1 Introduction to Artificial Intelligence
Since the invention of computers or machines, their capability to perform varied tasks went
on growing exponentially. Humans have developed the facility of pc systems in terms of their
numerous operating domains, their increasing speed, and reducing size with relevancy time. A
branch of engineering named Artificial Intelligence pursues making the computers or machines as
intelligent as kith and kin.
According to the daddy of Artificial Intelligence John McCarthy, it's “The science and
engineering of constructing intelligent machines, particularly intelligent pc programs”.
Artificial Intelligence (AI) may be a manner of constructing a computer, a computer-controlled
golem, or a computer code assume showing intelligence, within the similar manner the intelligent
humans assume.AI is accomplished by learning however human brain thinks, and the way humans
learn, decide, and work whereas attempting to unravel a tangle, then exploitation the outcomes of
this study as a basis of developing intelligent computer code and systems.
While exploiting the facility of the computer systems, the curiosity of human lead him to surprise,
“Can a machine assume and behave like humans do? “Thus, the event of AI started with the
intention of making similar intelligence in machines that we discover and regard high in humans.
The goals AI have:
 To produce knowledgeable Systems − the systems that exhibit intelligent behavior, learn,
demonstrate, explain, and recommendation its users.
 To Implement Human Intelligence in Machines − making systems that perceive, think,
learn, and behave like humans.

11. 2
Artificial intelligence may be a science and technology supported disciplines like Biology,
Psychology, Linguistics, Arithmetic, and Engineering. a significant thrust of AI is within the
development of computer functions related to human intelligence, like reasoning, learning, and
drawback resolution.
Out of the subsequent areas, one or multiple areas will contribute to build intelligent system.
 What is AI Technique?
In the globe, the information has some unwelcomed properties −
 Its volume is big.
 It isn't well-organized or well-formatted.
 It keeps ever-changing perpetually.
AI Technique may be a manner to prepare and use the information with efficiency in such the way
that −
 It should be perceivable by the people.
 It ought to be simply modifiable to correct errors.
 It ought to be helpful in several things though' it's incomplete or inaccurate.
AI techniques elevates the speed of execution of the complicated program which they are equipped
with.
Figure 1.1 – Artificial Intelligence different fields

12. 3
 History of AI:
1950 : Alan Turing thinks up of the Turing Test.
1951 : The SNARC is built. It is the first neutral net machine
1956 : The termed ‘Artificial Intelligence’ is coined
1963 : Machine learning theory is expounded
1964 : LISP program reads and solves word based algebra problems
1965 : Chatbot ELIZA is demonstrated
1972 : MYCIN diagnoses infectious blood disease
1975-1980 : ‘The Winter of AI’
1982 : Neutral Network theory gains popularity
1991 : DART logistics planning application used by US military
1997 : IBM’s DeepBlue beats world champion at chess
2000s : AI based algorithm begin to used in many markets
2005 : Introduction of web-based recommendations
2011 : IBM’s Watson wins US gameshow Jeopardy
2012 : Google Brain recognizes picture of cat
2014 : Google Brain describes the scene in picture
2015 : AI generalizes learnt information across different environments
2016 : DeepMind’s AlphaGO DNN beats Go champion
2017 : AI moves from cloud to the device with TensorFlow Lite, and Caffe 2 libraries
1.2 Artificial Intelligence in Retail
Behind the scenes of any AI-powered systems, there's a protracted and complicated
machine method involved the trained knowledge set for the rule to perform so the user gets an
awesome expertise. This expertise is therefore quick and seamless that the user thinks everything
is occurring as if by magic.
Here are some measures and fascinating facts regarding AI and retail:
 By 2020, 85% of client interactions in retail are managed by computing, as per Gartner.

13. 4
 According to Business corporate executive, shoppers UN agency move with on-line reviews
and opinions measures that 97% of a lot seemingly to convert with a distributor than customers
UN agency don't.
 70% folks millennials and 62% of millennials within the developed and developing countries
say they might appreciate a complete or distributor exploitation AI technology to indicate a
lot of fascinating merchandise.
Today’s e-commerce stores usually offers 2 general varieties of recommendations:- one is cross-
products recommendations from a distinct class and another is comparable product
recommendations (items you will like). For these recommendations, stores usually use ancient
cooperative filtering, cluster models, and search-based strategies to advocate the top user. These
recommendations don’t perform well most of the days. The other reasons they are not so effective
is because they are manually feed into the system and have inconsistent product tags. In fact, in
some cases of fashion products, these product tags are insufficient in describing a visually rich
fashion product
Modern technology advances have given Retailers access to exponentially a lot of knowledge
regarding what customers do and need. There are incredible chances for retailers to use analytics
to unlock the goldmine of data to cater to their customers. Retailers are now a days troubled to
convey customers a fascinating expertise with them. There are thoughts that simply developing
digital channels to access their account is enough to induce client loyalty. So as to produce a
fascinating expertise, it's essential what discourse data ought to run to their customers to form their
journey fruitful. For that, retailers should invest in an exceedingly technology stack which is able
to facilitate them to know what customers wish. AI and Machine learning can play an important
role, if not the crucial role to induce client intelligence. Computing (AI) permits retailers to ultra-
personalize the searching expertise at scale, exploitation vital volumes of information.
In this era of mass customization, customers became a lot of connected, a lot of tight and fewer
loyal. Straightforward availableness of comparable and competitive products/pricing totally
different retailers mean easier comparison and quicker switch between brands. The relationships
between customers and retailers has become temporary and mostly transactional. The demographic
trends have shivery implications for standard retailers because the millennials & younger
generations become the smallest amount loyal.

14. 5
 What & Why of Customer Intelligence:
Retailers got to adopt customer intelligence strategy as a result of it keeps the client at customer
of all operations. They need a deeper understanding of consumers through their purchase and
browsing history and transactions. Client Intelligence investments provides insights regarding the
client to grasp their persona.
It helps in building the client persona through that retailers will phase customers to enhance
cardinal and to own a targeted electronic communication. It helps them to interact at an emotional
level and strengthens the connection with the purchasers. Customer Intelligence is that the path to
true loyalty.
It’s vital for retailers to become an important partner for customers throughout their life. Which
implies turning customers into really Omni channel, supply ultra-personalized care, give a
compelling searching eco-system.

15. 6
CHAPTER 2
LITERATURE REVIEW
2.1 Different Types of AI
 Weak AI, which is also referred as Narrow AI, which mainly focuses on single task.
There is no genuine intelligence or self-awareness in case of a weak AI. iOS Siri is a
good example of a weak AI where several weak AI techniques are combine to function.
 Strong AI, which is also referred to as True AI, which is a computer that is as smart
as the human brain. This sort of AI usually will be able to perform all tasks that a
human could do. Robots are good example of Strong AI.
2.2 Basic Components of Artificial Intelligence
Figure 2.1 – Basic Components of Artificial Intelligence

16. 7
 Machine Learning | Learning from experience
Machine learning (ML) is an application of AI that provides computer systems with the ability to
automatically learn and improve from the experiences without being explicitly programmed. ML
focuses on the development of algorithms that can analyze the data and make predictions on its
own.
For example, ML can be used to predict what Netflix movies you might like, or the best route for
your Uber, machine learning is being applied to healthcare, life sciences, and pharma industries to
aid disease diagnosis, accelerate drug development and medical image interpretation.
How we Teach Machines?
Figure 2.2 – Machine Learning Process
For example, here we’ll pretend that we have a large data set of different songs in Spanish and
English. We want our model to be able to categorize the songs according to the given language.
Before diving into training, there are a few things that we need to check.
First, we need to make sure that we have high-quality data for processing. For all machine learning,
the data must be clean and organized, with no duplicates or irrelevancy. In this example, useful

17. 8
tags for each song could include the record label and artist name. These provide the AI with some
helpful clues when it makes predictions for using the training data.
Once our data is completely ready, it should be randomly assigned into three different categories:
training, validation, and testing.
Phase 1: Training
 Firstly by using the random variables available to us in the data, we will ask the model to
predict whether the songs are Spanish or English. This first time around, the AI has a little
idea of how any of the variables relate to its target, so this is no cause for alarm.
 Once the model has run the training data, we are able to start adjusting the parameters of
the variables in a way that we think will help the AI do better job next time. In our model,
perhaps we’re able to tweak things so that the algorithm can recognize sounds that are
present only in Spanish or English. We will have to spend a bit of time honing these
variables until we are ready to run the training data again.
 Now the model runs the training data again and does slightly better this time. Then we
simply repeat this process, improving the algorithm little by little each time when it
attempts to predict the language of our songs. Each one of these step cycles is called a
training step. During the first few attempts it will perform poor, but after a while we have
a machine that is ready for validation.
Phase 2: Validation
 It is the time to test our model against some new data. We take the validation data, with its
inputs and targets, and use it to run on our program. The algorithm will do better than when
it encountered the training data for the first time. Perhaps the algorithm has identified a few
songs correctly while others are still not identified.
 We again looks at our results and evaluate them. It’s possible we may see evidence of over
fitting, where the model has been trained a little too specifically to only recognize examples
which are present in our training data. Here the machine may be struggling to identify
words that sound similar in Spanish and English, such as music and música. We will need
to account for this in the next training step.

18. 9
 We will again go back to training with new variables in mind, adjusting and improving the
same algorithm. Once our model has done really well at categorizing the new songs in
validation steps, then we will skip straight to testing.
Phase 3: Testing
 Once the model has aced the validation process, it is ready for testing against data without
tags or targets. This simulates the state of the data where the algorithm will be expected to
perform against in the real world. If the model does well here, it is ready to be used for the
purpose it has been designed for. If it does not, then we have to go back to training until
we are satisfied.
 Deep Learning | Self-educating machines
Deep learning is a subset of ML that employs artificial neural networks that learn by processing
data. Artificial neural networks mimics the biological neural networks in the human brain.
Multiple layers of computer neural networks work together to determine a single output from many
inputs. For example, identifying the image of a face from mosaic of tiles. The machines learn
through positive and negative reinforcement of the tasks when they carry out the process, which
requires constant processing and reinforcement to progress.
 Natural Language Processing (NLP) | Understanding the language
Natural Language Processing (NLP), allows the computers to interpret, recognize, and produce
human language and speech. The ultimate goal of NLP is to enable effective interaction with the
machines we use every day by teaching the systems to understand human language in context and
produce the logical responses.
Real-world examples of NLP include Skype Translator, which interprets the speech of multiple
languages in real time to facilitate communication for ease of understanding.
 Neural Network | Making associations

19. 10
Neural networks enables the deep learning. Neural networks are the computer systems modeled
after neural connections in the human brain. The artificial equivalent of a human neuron is the
perceptron. Just like bundles of neurons create neural networks in the brain, the stacks of
perceptions creates the artificial neural networks in computer systems.
Neural networks learn by process training examples. The best examples come in the form of large
data sets which is the set of 1,000 cat photos. By processing this images (inputs) the machine is
able to produce a single output, answering the question, that “Whether Is the image a cat or not?”
This process analyzes the data many times for finding the associations and give meaning to
previously undefined data. Through various learning models, like the positive reinforcement, the
machine is taught it has successfully identified the object.
 Cognitive Computing | Making inferences from context
Cognitive computing is another important component of AI. Its purpose is to imitate and improve
interaction between the humans and the machines. Cognitive computing seeks to recreate the
human thought process in a computer model by understanding human language and the meaning
of images.
 Computer Vision | Understanding images
Computer vision is a technique that implements the deep learning into system and pattern
identification to interpret the content of images which includes graphs, tables, and pictures within
PDF documents, as well as text and video. Computer vision is an integral field of AI, enabling the
computers to identify, process and interpret data visually.
Applications of Computer Vision technology have already begun to revolutionize industries like
R&D and healthcare. Computer Vision is being used to diagnose patients faster by using Computer
Vision and machine learning by evaluating patients’ x-ray scans.

20. 11
CHAPTER 3
INDUSTRY PROFILE AND COMPANY PROFILE
3.1 Retail Industry
The Retail trade was valued at USD 23,460 billion in 2017 and is predicted to register a
CAGR of 5.3% throughout the forecast amount (2018 - 2023), to achieve USD 31,880.8 billion by
2023.The market provides merchandise like food, apparel, furniture, jewelry, and varied others.
Aside from this, the stores will be classified into the store, specialty merchant, web selling, and
varied others.
The retail market is mature and extremely competitive within the developed economies of Europe
and North America. On the opposite hand, the developing economies of Asia-Pacific, geographic
area, and Latin America are instrumental in driving the market growth. Client disbursal, which
usually accounts for around simple fraction of the GDP, has been a key indicator of the health of
the retail market. Moreover, the increasing strength of on-line looking has been a significant driver.
Aside from this, the growing smartphone penetration across countries is driving the e-commerce
channel. Also, web of Things (IoT) is reshaping the retail trade. It’s being deployed to
revolutionize the trade. However, value variation between on-line and brick & mortar stores will
challenge the retail market growth.
 Internet Retailing the Fastest Growing Segment of Retail Industry
Internet marketing is that the trendy approach of looking. With growing penetration of
smartphones and mobile devices and therefore the net services, e-commerce has emerged as a
serious looking platform within the world. Though the sector’s market size tripled over the past 3
business years, net marketing accounted for a mere 1.5% of the full retail sales in most of the
countries everywhere the globe. Mobile-first sites, dedicated apps, rising payment ways, and
different tools are making shopping on smartphones abundant easier.

21. 12
Table 3.1 – Global Retail Development Index, 2017
Figure 3.1 – Smart Phone Penetration by Country, 2016

22. 13
 India Retail Industry
Introduction
The Indian retail trade has emerged in concert of the foremost dynamic and fast industries because
of the entry of many new players. Total consumption expenditure is predicted to succeed in nearly
US$ 3,600 billion by 2020 from US$ 1,824 billion in 2017. It accounts for over ten per cent of the
country’s Gross Domestic Product (GDP) and around eight per cent of the use. Asian country is
that the world’s fifth-largest world destination within the retail area.
Market Size
India’s retail market is predicted to extend by sixty per cent to succeed in US$ 1.1 trillion by 2020,
on the rear of things like rising incomes and mode changes by socio-economic class and magnified
digital property. On-line retail sales are forecasted to grow at the speed of thirty one per cent year-
on-year to succeed in US$ 32.70 billion in 2018. India is predicted to become the world’s
quickest growing e-commerce market, driven by strong investment within the sector and fast
increase within the variety of web users. Varied agencies have high expectations regarding growth
of Indian e-commerce markets. Luxury market of
India is predicted to grow to US$ 30 billion by the tip of 2018 from US$ 23.8 billion by growing
exposure of international brands amongst Indian youth and better getting power of the class in tier
2 and tier 3cities, per Assocham.
Investment Scenario
The Indian retail commerce has received Foreign Direct Investment (FDI) equity inflows totaling
US$ 1.42 billion throughout Apr 2000–June 2018. With the rising would like for trade goods in
several sectors as well as client physics and residential appliances, several corporations have
invested with within the Indian retail area within the past few months.
 Beccos, a South Korean designer is ready to enter the Indian market with an investment of
about Rs 1 billion (US$ 14.25 million) and open fifty stores by Gregorian calendar month
2019.
 Walmart Investments Cooperative U.A has invested with Rs 2.75 billion (US$ 37.68 million)
in Wal-Mart Asian country Pvt Ltd.

23. 14
Government Initiatives
The Government of India has taken varied initiatives to enhance the retail business in India. A
number of them area unit listed below:
 The Government of India might amendment the Foreign Direct Investment (FDI) rules in
food process, in a very bid to allow e-commerce corporations and foreign retailers to sell
created in India shopper merchandise.
 Government of India has allowed 100% Foreign Direct Investment (FDI) in on-line retail of
products and services through the automated route, thereby providing clarity on the present
businesses of e-commerce corporations operative India.
Road Ahead
E-commerce is increasing steady within the country. Customers have the ever increasing
alternative of merchandise at the bottom rates. E-commerce is maybe making the largest revolution
within the retail business, and this trend would continue within the years to return. India's e-
commerce business is forecasted to achieve US$ 53 billion by 2018. Retailers ought to leverage
the digital retail channels (e-commerce), which might alter them to pay less cash on realty whereas
reaching bent on a lot of customers in tier-2 and tier-3 cities.

24. 15
3.2 AI Industry
Figure 3.2 - Revenues from the artificial intelligence (AI) market worldwide from 2016
to 2025 (in million U.S. dollars)
 Artificial Intelligence (AI) Market Overview
The global computing market size is anticipated to succeed in $169,411.8 million in 2025, from
$4,065.0 million in 2016 growing at a CAGR of 55.6% from 2018 to 2025. Computing has been
one in all the fastest-growing technologies in recent years. AI is associated to human intelligence
with similar characteristics like language understanding, reasoning, learning, drawback
determination, and others. Makers within the market witness huge underlying intellectual
challenges within the development and revision of such a technology. AI is positioned at the core
of consecutive info software system technologies within the market. Corporations like Google,
IBM, Microsoft, and different leading players have actively enforced AI as an important a part of
their technologies.
The AI market is metameric by technology, trade vertical, and earth science. The assorted
technologies are sub-divided into machine learning, tongue process, image process, and speech
recognition. In 2016, the machine learning phase dominated the market, in terms of revenue, and

25. 16
is anticipated to keep up this trend within the returning years, as a result of increase in demand for
computing trade solutions. Supported trade verticals, the market is categorized into media &
advertising, retail, medium & IT, healthcare, automotive & transportation, etc. The IT phase is
anticipated to dominate the worldwide computing artificial intelligence market throughout the
forecast amount. Geographically, market is analyzed across Asia-Pacific, North America, Europe,
and LAMEA. In 2017, North America region contributed the best share within the computing
market and is anticipated to secure the leading position throughout the forecast amount, as a result
of the presence of key corporations and enormous investment within the AI market.
Figure 3.3 – Current Adoption and Future demand of AI
3.3 Company Profile - LENSKART
Lenskart Solutions Pvt Ltd operates as an online shopping portal for men and women eyewear in
India. It provides sunglasses, goggles, frames, contact lenses and as well as sunglasses and frames
for kids. The company also owns and operates eyewear retail stores in India; and provides home
eye check-up services in Mumbai, Pune, Delhi, Bengaluru, Chennai, Hyderabad, Kolkata,
Gurugram, Noida, Faridabad, and Ghaziabad. Lenskart was formerly known as ‘Valyoo

26. 17
Technologies Pvt Ltd’ and changed its name to Lenskart Solutions Pvt Ltd in May 2015. The
company was incorporated in 2008 and is headquartered in New Delhi, India with retail stores in
Delhi, Visakhapatnam, Hyderabad, Agartala, Nagpur, , Kolhapur, Bengaluru, Sangli, Mumbai,
Pune, Bhubaneshawar, Siliguri, Ahmedabad, Mangalore Haridwar, Trivandrum, Kochi, Chennai,
Tirupati, Raipur, Lucknow, Faridabad, Meerut, Erode, Ludhiana, Jamshedpur, Kolkata, Varanasi,
Surat, Ranchi, Cuttack, Kannur, Dhanabad, Indore, and Madurai and Goa.
Being a provider of an online optical store, Lenskart is intended to provide a various range of eye
wear in India. The company's online optical store sells prescription eye-wear, sunglasses, branded
contact lenses and accessories, enabling customers to get the latest eye-wear collections with free
home delivery guarantee and 365 days return policy.
Website: www.lenskart.com
Primary Industry: Accessories
Figure 3.4 - Key Highlights of Lenskart

27. 18
CHAPTER 4
RESEARCH METHODOLOGY
4.1 Data Collection Methods
 Primary Data: - Primary data is the data that are collected by different techniques like
questionnaire, Depth interview, Survey, Schedules etc. In this project, primary data has
been collected by the means of questionnaire through Google forms.
 Secondary Data: - Secondary data is the data that are already available i.e.: they refer to
the data which have already been collected and analyzed by someone else. Secondary
Data has been used in this Project by the help of articles and research papers.
 Research Design
Research design is mainly classified into three types as:-
 Exploratory Research Design
 Descriptive Research Design
 Causal Research Design
 Research Design Used
Descriptive research studies are the studies which are mainly concerned with described the
characteristics of particular individual. In descriptive, the researcher must be able to define clearly,
what he wants to measure and must find adequate methods for measuring it along with a clear cut
definition of population he want to study. Since the aim of the study is to obtain complete and
accurate information in the said studies, the procedure to be used must be carefully planned. The
research design must always make enough provision for maximize reliability and protection
against the bias, with due concern for the economical completion of the research study.

28. 19
4.2 Sample Design and Techniques
A Sample Design is a plan for obtaining a sample from a given population. It refers to the technique
that must be adopted in selecting items for the sampling designs are as below:
 Sample Size:-
The sample size has been 81 respondents.
 Sampling Method
In this research project, I am using Convenience Sampling Method.
 Sample Type
The Area Sampling is Maharashtra region.

29. 20
CHAPTER 5
DATA ANALYSIS AND HYPOTHESIS TESTING
5.1 Data Analysis
The survey has been done to know the impact of Artificial Intelligence on enhancing the user
experience with the help of LENSKART 3D Try On feature
 3D Try On feature
The 3D Try On feature records the user face from multiple angles letting it map the face and then
when the users try the frame on their face virtually, customers can swipe on the image to turn the
head to the left and right as well, to get a view of the glasses from different angles. This online 3D
face modelling trial provides preferences and historical data to make frame selection faster,
effective and fun for the buyers.
Figure 5.1 – Lenskart 3D Try On Feature
The following questions were asked to the respondents to know their views:
Google Form:

30. 21

31. 22

32. 23
 Data Analysis process
Once the data has been collected, the next task is to aggregate all the data in a meaningful manner.
A table was prepared to bring out the main characteristics of the data. The researcher should have
a well thought of all the framework for processing and analyzing data and this action should be
done prior to the collection.
It includes the following activities:
I. Editing
The first task in data processing is always the editing. Editing is the process of examining errors
and omissions in the collected data and making necessary corrections in the same (e.g. not needed
responses).
II. Tabulation
Tabulation comprises of sorting the data into different categories and counting the number of cases
that belong to each category.
III. Analysis
After all the above three steps, the most important step is analysis of the data.
The following is the analysis of the data collected through survey:
(1)

33. 24
(2)
(3)
(4)
1%
67%
10%
22%
<18
18-30
31-45
45<

34. 25
(5)
(6)
(7)

35. 26
(8)
(9)
(10)

36. 27
5.2 Hypothesis testing
Hypothesis 1A: Due to the use of 3D Try On (A.I Feature) the sales of the Lenskart increases.
Hypothesis 1B: Due to the use of 3D Try On (A.I Feature) the sales of the Lenskart doesn’t
increases.
Hypothesis 2A: Due to the use of 3D Try On (A.I Feature) of Lenskart, the customer experience
is improved.
Hypothesis 2B: Due to the use of 3D Try On (A.I Feature) of Lenskart, the customer experience
doesn’t improved.
 For Hypothesis 1
Google search trends (visits per 100) for Lenskart website during 2014 to 2015:
Figure 5.2 - Google Trends for Lenskart
As the 3D Try On feature got introduced in first half of 2015, after that the search trend of the
brand from viewers got increase. Thus there is more possibility that the sales of the eye wears
increased due this 3D Try On (Artificial Intelligence feature).

37. 28
Figure 5.3 – Top Competitors of Lenskart
Lenskart Coolwinks GKB Opticals Titan Eyeplus
Estimated visits 2.2 M 1.1 M 42.7 K 433.6 K
Time on Site 4.01 mins 2.49 mins 2.16 mins 2.46 mins
Page Views 4.82 4.56 2.78 3.35
Bounce Rate 43.98% 61.85% 64.54% 62.54%
Table 5.1 - Website Engagement (Online Eyewear Retailers)
In Online Eyewear Retailers, only Lenskart provides 3D Try On Feature. From above, Lenskart
has better customer engagement on its website than other. These can be due to the unique Artificial
Intelligence feature provided by Lenskart. As from above the estimated visits of Lenskart is the
highest among all which in turn gives more sales converting from visits.
Thus from here, we can conclude that this A.I feature of 3D Try On can be very helpful for business
for improving sales performance.
Year Revenue(in Crores) Percentage of Growth
2014-2015 132.9 -
2015-2016 143.2 7.75%
2016-2017 179 25%
Table 5.2 - Revenue Information of Lenskart

38. 29
The sales growth has been more in FY16-17 than FY15-16.The 3D Try On feature has been
introduced in first half of 2015.Thus,we can say that due this feature and also offline expansion of
Lenskart the sales has been increased.
Thus from above, Hypothesis 1A can be proved i.e. Due to the use of 3D Try On (A.I Feature) the
sales of the Lenskart increases.
 For Hypothesis 2
As from the survey conducted,
 81.9% of respondents have given rating greater than equal to 5 for the usefulness of 3D
Try On feature.
 More than 35% of respondents said that 3D Try On feature provides better user
experience than conventional product images.
 85% of respondents have said that 3D Try On feature supports their buying decision with
ratings greater than equal to 5.
 63.3% of respondents have said that 3D Scanning feature should be added while buying
clothes online.
 71.8% of respondents have said that 3D Scanning feature should be added in apparel
stores while buying clothes.
Thus from above, Hypothesis 2A can be proved i.e. Due to the use of 3D Try On (A.I Feature) of
Lenskart, the customer experience is improved.
 Limitation of the Study
 The Sample Area has been limited due to time constraints.
 Users were unlikely to change the Response reaction as the survey was conducted on
Google Forms.

39. 30
CHAPTER 6
FINDINGS AND RECOMMENDATIONS
6.1 Findings
With the use of Artificial Intelligence in future, the retailers in India can increase their sales,
improve business performance and also can able to improve the customer experience of buying
items both online(i.e. internet) and offline(i.e. local stores).
6.2 Recommendations
The applications of Artificial Intelligence in Retail for enhancing the customer experiences is
what should we look forward. Below are the applications which can be implemented in Indian
local retail stores or retail websites for improving customer experience and also retailers
business.
 Applications of Artificial Intelligence in Retail
Amazon was one among the primary to use the potential of computing (AI) within the early 2000s,
and 35 % of the company's sales were before long attributed to those AI innovations. Since then,
AI technology has evolved at associate degree astronomical rate. Computing will serve retailers
with good recommendation engines, alter most processes, replace human labor for explicit tasks,
predict demand, and by doing this stuff increase overall productivity and revenue.

40. 31
Figure 5.4 – What AI can bring to Retail

41. 32
#1 Virtual fitting rooms and mirrors
Among applications of Artificial Intelligence in retail we need to mention virtual dressing rooms
and mirrors. The virtual fitting room is of a great help for busy shoppers as they can try out find
the right outfit, accessories and manifold apparel that perfectly matches with each other and do all
in very less time.
For example, Moda Polso lets its customers to create their own avatars. These virtual avatars let
shoppers try on an unlimited number of outfit options and helps to make a purchase decision easily
and quickly.
Figure 5.5 - Moda Polso Customers experiment
In above image, Moda Polso Customers experiment with different garments on their own image
via the touchscreen.
A Canada-based tech startup named Me-Ality, has developed a virtual fitting kiosk that can scan
a shoppers complete body. A scan takes about 20 seconds of time and measures 200,000 different
points on the body. Brooks Brothers, Gap, J.Crew, Levi’s,American Eagle and Old Navy has set
these scanners in their stores and after that they saw dramatic increase in their sales.Virtual
dressing rooms are very important for online stores, as online shoppers return 25% of their clothing

42. 33
they buy, with 70% of returns because of the wrong size. Specsavers was one among the first
retailer to offer a Virtual Try On feature. With Specsavers, a customer can scan their face with
the camera on their mobile, tablet or desktop and virtually try on glasses in one click.
#2 Visual product search
Artificial intelligence has opened visual search to retailers and also allows customers to upload
images and find similar or identical products. AI-powered technology examines an image and
analyzes its shapes, patterns and colors to identify an item.
For example, Cortexica, a London-based Artificial Intelligence company, has developed image
recognition technology that promises 95% of accuracy.
Another Example of Visual Product Search is American Eagle Outfitters also offers visual search
in its mobile application. American Eagle image recognition technology allows the customers to
not only find the similar or exact clothing but also get recommendations on what looks well with
it.
Figure 5.6 – Image recognition technology on mobile

43. 34
In above Image, Image recognition technology detects a women’s black leather jacket in the image
and suggests accessories on the right.
#3 In-store assistance
Many retailers has invest in Artificial Intelligence-driven technologies that can both assist
customers while shopping in their physical stores and help their staff to handle customer inquiries
effectively. John Lewis spent £4 million in 2017 on shop floor application for personnel use of
customer. This application equips the employees with information about products and stock
availability. Thus, employees are able to assist and give right answers to questions asked by the
customers.
For example, Kroger company wants to roll out smart shelves in almost 200 stores by the end of
2018.This company’s technology will replace paper price tags with instantly changing digital tags.
In addition to prices, smart shelf tags will display nutritional data, product details, current
promotions, and video ads.
Lowe’s company has employed AI to develop an autonomous in-store robot called Lowebot.
Lowebot helps to navigate the customers through stores and find goods in multiple languages. Due
to real-time monitoring capabilities, the robot helps the company to manage its inventory
efficiently.
Figure 5.7 – LoweBot Autonomous Robot

44. 35
In above image, the LoweBot has an interactive screen that customers can use to search for
items.
#4 Eliminating overstocks and out-of-stocks
Аn excess or short offer of merchandise have an effect on your company’s gain and prices retailers
worldwide by $1.1 trillion every year. Leftover stock is commonly marked down and results in
low sales turnover. Out-of-stock things, on the opposite hand, work lost sales and discontent
customers will lead to simply switch to competitors.
One of the business applications of AI in retail is restocking. AI helps retailers fill again provides
by characteristic demand for a selected product supported by
 sales history
 location
 weather
 promotions
 trends
This way corporations will forestall underperforming product from buildup, stock what customers
area unit seemingly to shop for, come through quicker deliveries, cut back returns, and save several
cash.
Figure 5.8 – Usage of AI in Supply Chain Management

45. 36
For example, H&M uses AI to investigate store returns, receipts, and loyalty cards to predict future
demand for attire and accessories and manage inventory.
Morrisons has partnered with BlueYonder, a number one AI solutions supplier for retailers, to
optimize stock prognostication and filling across its 491 stores. Use of computing (AI) has helped
the corporate to scale back shelf gaps in-store by up to 30%.
#5 Product categorization
Artificial intelligence is smart way to classify merchandise products. LovetheSales.com employs
machine learning to categorized almost 1,000,000 commodities from varied retailers. The
algorithmic rule tags merchandise and classifies them into product classes and shows to customers
which they were looking for.
Lalafo has created a breakthrough by sorting merchandise and services via AI-powered image
recognition. Once sellers wish to promote product on Lalafo, they'll simply transfer a picture of
those product with no added description. Computer vision and machine learning acknowledge
things in photos, place these things within the applicable product class, and suggests there prices.
The platform presently processes over 900 requests per second. AI has helped Lalafo increase
content relevance and improve sales.
#6 Chatbots for customer support
Chatbots another popular application of AI within the retail trade. Chatbots facilitate retailers to
supply great customer service, helps the customer to find items they want on the site, inform them
about concerning new collections, and provide them with various apparel similar to which they’ve
already chosen. For instance, if a customer has already additional black jeans to the cart, a chatbot
can give them new silver Converse shoes to finish the same look.
80% of brands globally are using or planning to use chatbots. Burberry and Tommy Hilfiger have
already launched AI-driven bots on Facebook messenger that guide customers through their latest
collections and answer to their inquiries.

46. 37
1-800 Flowers has additionally launched an AI-powered platform named Celtic deity (Gifts when
you Need). Celtic deity emulates electronic messaging platforms like WhatsApp and replies to
client queries successfully, facilitate customers to realize the most preferable gift options, and
assist them through the complete searching expertise. The corporate uses AI to reach to every
customer with individual tailored offer.
Figure 5.9 – GWYN Chatbot
In above Image, GWYN aims to help last-minute on-the-go shoppers
#7 Voice product search
Artificial intelligence has made it possible for voice search of any products. Many prominent
brands such as Costco, Kohl’s, Tesco,Target and Walmart use either Amazon or Google AI
technology and smart devices to serve customers with fast and easy search.
Customers don’t need to type their queries on small devices anymore, they can just ask Alexa to
add potatoes or new compact powder by MAC to their shopping bag. Customers can also inquire
by using smart assistants to know about the estimated arrival time and current status of delivery or
reorder items they already bought.

47. 38
#8 Cashier-free stores
The robotization of stores helps reduce the number of employees needed by diminishing the
lines and dramatically save operational costs for a company.
For example, Amazon AI has enabled checkout free stores. Just Walk Out Shopping technology
along with Amazon Go app automatically react to a customer taking from and returning
products to the shelf. Just Walk Out Shopping technology records the products which are taken
on the customer virtual card. Once a customer is done shopping, they can just leave the Amazon
shop. Later, the customer receives a receipt and the company will charge their Amazon account.
Amazon plans to create more such Artificial Intelligence driven supermarkets like Amazon Go
with a mere just 6 to 20 employees at any shop.
Figure 5.10 – Amazon Go Store
In above Image, it’s an Amazon Go Store where Customers can buy their items

48. 39
CONCLUSION
I was an under-graduate level student. I am so happy that, I could complete the last
phase of my postgraduate life doing research on Artificial Intelligence in Retail. In this span, it
can be said that the knowledge gained from working has helped to open up sea of opportunities
to explore and understand in future. It has also helped in understanding the how technology will
evolve in future. It helped me to gain ideas and to recommend or come up with innovative ideas
as well. Overall, my Capstone project has been a great success. This has been a great learning
experience for my career.
I would like to once again appreciate everyone who has made my project a superb experience.
Again, this report is done with a lot of limitations and obstacles. Thanks to so many people who
helped me doing this report.

49. 40
BIBLIOGRAPHY
Books
Thomas Laville (2017), Artificial Intelligence: Guide for Absolute Beginner
Newspapers and Magzines
Abhishek Law,The Hindu Business Line ‘Eye-wear e-tailer Lenskart looks at 150% growth this
fiscal’, 22 January 2018.
Websites
Imran Shaikh,(13th
Jan,2019) “UX audit of Lenskart 3D Try On and DITTO™”, accessed on
12/03/2019 at https://uxplanet.org/ux-audit-of-lenskart-3d-try-on-and-ditto-e20511dd27e8.
Sumangala Varun,(21st
Oct,2015) “Lenskart launches online trial room feature to enhance
customer experience”, accessed on 15/03/2019 at
https://indianonlineseller.com/2015/10/lenskart-launches-online-trial-room-feature-to-enhance-
customer-experience/.
Rozelle laha (18th
Dec, 2018), “Lenskart: A 6/6 vision for growth”,accessed on 19/03/2019 at
https://www.fortuneindia.com/enterprise/lenskart-with-an-eye-on-growth-2/102790