Artificial Intelligence and Machine Learning in Business Management

Concepts, Challenges, and Case Studies

Sandeep Kumar Panda, Vaibhav Mishra, R. Balamurali, Ahmed A. Elngar, Sandeep Kumar Panda, Vaibhav Mishra, R. Balamurali, Ahmed A. Elngar

• Publisher: CRC Press
• Year: 2021
• Print ISBN:
  9780367645557
• eBook ISBN:
  9781000432145
• Topic: Business
• Subtopic: Information Management
• Index: Business

Artificial Intelligence and Machine Learning in Business Management

The focus of this book is to introduce artificial intelligence (AI) and machine learning (ML) technologies into the context of business management. The book gives insights into the implementation and impact of AI and ML to business leaders, managers, technology developers, and implementers.

With the maturing use of AI or ML in the field of business intelligence, this book examines several projects with innovative uses of AI beyond data organization and access. It follows the Predictive Modeling Toolkit for providing new insight on how to use improved AI tools in the field of business. It explores cultural heritage values and risk assessments for mitigation and conservation and discusses on-shore and off-shore technological capabilities with spatial tools for addressing marketing and retail strategies, and insurance and healthcare systems.

Taking a multidisciplinary approach for using AI, this book provides a single comprehensive reference resource for undergraduate, graduate, business professionals, and related disciplines.

1 Artificial Intelligence in Marketing

Rachita Kashyap and Rishi Dwesar

Marketing and Strategy Department, IBS, IFHE University, Hyderabad

DOI: 10.1201/9781003125129-1

Contents

• 1.1 Introduction
• 1.2 AI, ML and Data Science
• 1.3 AI and Marketing
• 1.4 Benefits and Detriments of Using AI in Marketing
  • 1.4.1 Benefits
  • 1.4.2 Detriments
    • 1.4.2.1 Amazon Go (Caselet)
    • 1.4.2.2 Technical Working of Amazon Go
    • 1.4.2.3 Issues Related to Amazon Go Technology
• 1.5 Marketing Plan and AI’s Potential
• 1.6 Future
• References

1.1 Introduction

During the early days of AI in the 1950s, scientists were asking questions like “Can machines think?” and were looking into deep, complex issues of mechanizing emotional intelligence. However, in today’s world of Amazon’s Alexa, Apple’s Siri, Google Assistant etc. machines have become far more capable, yet AI still has a long way to go. In simple terms, AI can be explained as human intelligence exhibited by machines and can be broadly classified into Artificial Narrow Intelligence (e.g., smart speaker, self-driving car, AI in farming and factories) and Artificial General Intelligence (machines can do anything a human can do). The benchmark for AI is to be as good as human intelligence, and to possess ability to reason, see and communicate like humans do. Though capability of AI has improved by leaps and bounds over the years, it is still far off from that benchmark in comparison to human intelligence. Nevertheless, recent advancements in the field of AI reflect a very promising future.

AI is Machine Learning (ML) driven. ML deals with training a computer to perform specific tasks and functions automatically. Usually, these tasks are exhaustive, repetitive and often too complex for humans to do efficiently. Machines can learn through supervised, unsupervised and reinforced learning. In supervised learning, several input and output sets are provided to the machine. Through this approach, data is fed to an algorithm and the machine tries to recognize the relationship between the input and the output. When the machine has stopped learning – in other words, has learned optimally – the learned model can predict the value or the class of new data points. For example, a system can be trained to differentiate between a kangaroo and a koala. By feeding the system with dozens of images of both animals, the system will learn about the features which distinguish each and so improve its prediction. In unsupervised learning, data is analyzed without trying to make any predictions. It is focused on learning and understanding the underlying structural properties and associations in between the observed data. This kind of learning can be used in detecting outliers, classifying and segmenting customers and the market. In reinforcement ML, the system does not have historical data to draw conclusions upon; instead, the algorithm learns by taking different actions and evaluating their successes and failures. Reinforcement learning is used by Facebook in advertising on its platform. The system tests the advertisements on full spectrum when it is flighted for the first time. With time, and when sales rise, Facebook’s algorithms analyze the data available, and it then shows the advertisements to certain sets of customers, in certain geographical locations, at certain times of the day and using certain on-screen placements.

1.2 AI, ML and Data Science

We often see AI, ML and data science in use together, and they are considered to be lucrative career options today. Data science can be defined as a broad field of study which pertains to data systems and processes which are aimed at maintaining datasets and deriving meaning from them. With the advent of technology and the Internet, today almost all organizations generate a large volume of data through their daily transactions, and it becomes problematic for these organizations to monitor, store, organize and extract important information from this data. Data scientists use a combination of tools, algorithms, applications and principles to extract useful information from various random data clusters, and then use this information to guide business processes to reach organization goals. The information extracted can be used to study ongoing data trends in any field of business and is helpful in presenting business forecasts and setting courses of action based on insights found and inferences made. The best example of ML is Netflix suggesting movies to the customers based on their movie-viewing behaviour and Amazon recommending books based on the past purchases of customers on the website. With the help of ML, marketers can provide customers with customized content as well as suggest other products that they may wish to purchase.

ML is a field of study that gives computers an ability to learn without being explicitly programmed, whereas data science deals with extracting knowledge from data. Deep learning is a big artificial neural network which mimics the network of neurons in a human brain. It is a subset of ML and is called “deep learning” because it uses deep neural networks for learning. The machine uses different layers to learn from the data and the depth of the model is represented by the number of layers in the model. Deep learning is a new term used within the world of AI. Refer to Figure 1.1 Relationship between AI, ML and Deep Learning.

Figure 1.1 Relationship between AI, ML and data science.

The relationship between AI, ML and data science is given in Figure 1.1.

1.3 AI and Marketing

In one century, human civilization has developed by leaps and bounds in terms of technology, healthcare, economy and in every possible materialistic dimension. In the years to come we are likely to witness replacement of human-driven cars to self-driving cars, doctors operating remotely through robotic surgical devices, nanotech self-cleaning clothing, 3D printers facilitating instant delivery of goods, meaning manufacturing time can be brought down, implantable communication devices replacing mobile phones, and many other technical advancements. It would come as little surprise if all advertising and marketing tasks were to be managed wholly by computational systems. Currently, marketers use a lot of technology, but in the coming decades, AI and ML methods will take the marketing game to a new level.

Practitioners and academicians have anticipated that AI will change marketing strategies and customer behaviours. According to a survey conducted by Salesforce, AI will be the technology most adopted by marketers in the coming years. With the help of AI, business processes are being automated and so machines are then able to perform preset tasks with higher accuracy and far less human intervention, such as transferring data, sending promotional mail to existing as well as potential customers, updating customer files, replacing lost ATM cards, reading documents to extract key points using natural language processing etc. With the help of AI, companies can gain insights from the vast amount of transaction and customer data which will include not only numeric data type, but also text, images, audio recordings of customers conversing with the customer care service provider, facial expressions, and even voice tones. By employing AI in daily functioning, companies can better predict customer choices, deploy appropriate digital marketing strategies or anticipate potential credit fraud.

Marketing bots are one of the most popular forms of automation right now. A bot is basically a piece of software which can be programmed to carry out a specific set of actions on its own. Bots are usually cheap to setup, and easy to program and run. Bot-powered commerce is our modern-day manifested destiny and is the future of marketing. For example, in order to purchase a bulb for your newly bought reading lamp, you need to visit different websites, scroll through a number of pages, fill out forms regarding your shipping address, give payment information and so on. But if there is a bot, you just need to tell it to find a bulb for your reading lamp, and it will guide you through different bulb hues, voltages, etc. and then place the order for you. Behind the screen, the bot leads you through a concatenation of questions in order to better understand your intent, and deliver the right information to you. AI bots can provide both customer as well as sales support services, are available 24/7, have very low error rates, and their deployment can be scaled up or down according to demands.

Here are a few functions that a bot can perform which can be beneficial to any business:

1. Assist website visitors looking for answers about products.
2. Help in conducting marketing research.
3. Qualify leads.
4. Help in tracking individual team members’ work and keep the whole team updated with each other’s work.
5. Personalizing advertisements for customers.

With all the descriptions suggested above, AI offers the potential to reduce the costs incurred and increase revenues. Revenues can be increased by making informed and improved marketing decisions (e.g., product recommendations, competitive pricing, personalized promotion, enhancing customer engagement). The costs may decline due to automation of simple marketing tasks and free up human agents to handle more complex marketing tasks.

There is a misconception that AI is replacing humans in their jobs, but firms can use AI to amplify their employees’ capabilities. For example, Stitch Fix, a leading clothing and service provider, use AI bots in assisting their employees to provide a better service to their customers. With the help of AI, stylists identify the best clothing styles for their customers by integrating all the data provided by the customers while expressing their preferences, general style trends, handwritten notes, Pinterest boards and preferences of other customers in the same segment. Ginni Rometty (CEO of IBM), in his media interactions has often indicated that AI would not lead to a world of man “versus” machine but rather a world of man “plus” machines [1].

1.4 Benefits and Detriments of Using AI in Marketing

1.4.1 Benefits

1. Personalization and relevant messages
   AI fuelled predictive analytics can help companies by tapping the right customer base and analyzing their browsing history, and then showing appropriate advertisements to the right set of people. This can help companies understand their customer preferences better and then make appropriate recommendations. This is being used widely by Amazon and Netflix, saving billions by keeping customers hooked to their services and avoiding cancellation of services. As a marketer, AI gives you much power in terms of developing certain data points which lets you guide your customers to the right product.
2. Streamlining the marketing efforts
   Through deep learning AI can study consumer behaviour patterns and predict which segment of customers are likely to make a certain kind of purchase. This can help businesses in targeting the customer base more accurately without wasting time and money on less probable leads.
3. Cost saving
   According to various research surveys conducted worldwide, around 85% of the interactions between brands and customers are going to happen online. As compared to other advertising mediums like prime time tele advertisements, print advertisements, billboards etc. online advertising is cheaper as well as more precise in targeting the right set of customers with the aid of AI.

1.4.2 Detriments

1. Human control is still required
   AI cannot function without human intervention, as it lacks the creativity, flexibility and imagination which makes humans the epicentre of the marketing world. Humans have various tastes, preferences, experiences etc. which enable them to make better decisions than machines, which are run on algorithms comprising from formulas, statistics, commands etc.
2. Algorithms can be wrong
   Due to bad data AI can develop biases. For example, automatically preferring and shortlisting CVs of white males over people belonging to other ethnicities, genders, colours etc. because previously most of the successful people in those positions were white males. And this makes AI infer that white males are better suited for those positions over other people.
3. Automated chat boxes and machine answering calls is n...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Dedication
• Contents
• Preface
• Acknowledgements
• Contributors
• Editors
• 1 Artificial Intelligence in Marketing
• 2 Consumer Insights through Retail Analytics
• 3 Multi-Agent Paradigm for B2C E-Commerce
• 4 Artificial Intelligence and Machine Learning: Discovering New Ways of Doing Banking Business
• 5 Analysis and Comparison of Credit Card Fraud Detection Using Machine Learning
• 6 Artificial Intelligence for All: Machine Learning and Healthcare: Challenges and Perspectives in India
• 7 Demystifying the Capabilities of Machine Learning and Artificial Intelligence for Personalized Care
• 8 Artificial Intelligence and the 4th Industrial Revolution
• 9 AI-Based Evaluation to Assist Students Studying through Online Systems
• 10 Investigating Artificial Intelligence Usage for Revolution in E-Learning during COVID-19
• 11 Employee Churn Management Using AI
• 12 Machine Learning: Beginning of a New Era in the Dominance of Statistical Methods of Forecasting
• 13 Recurrent Neural Network-Based Long Short-Term Memory Deep Neural Network Model for Forex Prediction
• 14 Ethical Issues Surrounding AI Applications
• 15 Semantic Data Extraction Using Video Analysis: An AI Analytical Perspective
• Index