Artificial Intelligence (AI) and the Industry. A threat or an opportunity?

Table of Contents

Dedication

Abstract

Table of Contents

Preface

Chapter 1: Introduction
1.0 Overview
1.1 Background of Study
1.2 Statement of Research Problem:
1.3 Research questions:
1.4 Research Hypotheses:
1.5 Research objective:
1.6 Significance of the study:
1.7 Scope/Limitation of the study:
1.8 Definition of key terms:

Chapter 2: Literature Review
2.0 Overview
2.1 History of Artificial Intelligence
2.2 Major achievements of Artificial Intelligence in modern history
2.3 Investments in AI for the future
2.3.1 Robotic workforce
2.3.2 Ubiquitous Artificial Intelligence
2.3.3 Voice Assistants
2.3.4 Capsule Networks
2.3.5 AI medical diagnostics
2.3.6 Robotic Process Automation
2.3.7 Text Analytics and NLP
2.3.8 Natural Language Generation
2.3.9 Speech Recognition
2.3.10 Decision Management
2.3.11 Deep Learning Platforms
2.3.12 The arrival of “backprop”

Chapter 3: Improving Business Productivity with AI
3.0 Overview
3.1 Introduction
3.2 Adopting AI in Business: Benefits
3.2.1 Data Prediction and Analysis
3.2.2 AI and Customer Interactions
3.2.3 AI and Business Process Automation Benefits
3.3 Disadvantages of AI in Business
3.3.1 Cost
3.3.2 Technical Challenges
3.3.3 Privacy Issues
3.3.4 Unemployment and AI
3.4 An Application of Machine Learning in Business
3.4.1 ChatBot
3.4.2 Improving the business process with an AI ChatBot
3.4.3 Customer Demands
3.4.4 Drawbacks
3.5 AI ChatBot Design and Implementation
3.5.1 Business Requirement
3.5.2 System Architecture
3.5.3 Conversation Process
3.5.4 Dataflow Diagram
3.5.5 Conversation Flow
3.5.6 ChatBot Prototype 1 Demonstration

Chapter 4: Research Methodology
4.0 Overview
4.1 Research Design
4.2 Research Period
4.3 Characteristics of the Research Population
4.4 Survey Sampling Technique
4.5 Data Collection
4.6 Limitation of Research

Chapter 5: Data Presentation, Analysis, and Interpretation
5.0 Overview
5.1 Analyses and Presentation of Primary Data
5.2 Summary of findings

Chapter 6: Conclusion
6.0 Conclusion
6.1 Recommendations and Suggestion for further research

Bibliography

Dedication

This project is dedicated to my family and friends; with whom I've learnt that even the largest task can be accomplished if it is done one step at a time. This project is also dedicated to my supervisor and instructors who have been very supportive in various ways during this research.

Abstract

Technology remains a catalyst for change and rapid evolution in businesses today. Over the past decade, so much has evolved from the way we think, act and make daily decisions.

Over centuries, humans have played an important part in history and a huge role in shaping the world and society we live in today. This is because of the importance of our intelligence.

For years, we have not only stimulated creation and destruction, but also have tried to understand how we think, perceive, understand, predict, or manipulate a world beyond our comprehension. Today, technology through the form of artificial intelligence goes beyond this to not only understand human intelligence but to create the intelligent being.

All these changes have been influenced by technology. We have seen trends redefine logic and business practices. Today, businesses are ever more interested in this evolution compared to anytime in the past. It is for this reason that this research is conducted to study the current trend of artificial intelligence, a technology that is gaining popularity across various industries and businesses.

This research will look at the history of artificial intelligence, a few subfields within artificial intelligence, and more importantly an application of machine learning, as an approach towards improving business productivity.

Keywords; Artificial intelligence, machine learning, productivity, technology, data, algorithm, business process.

Preface

This is a master's degree thesis project. The aim of this research is to identify ways of improving business productivity with artificial intelligence. This research is subdivided into six sections.

Section one covers the introduction of this research work. It gives a summary of the objectives, problem, hypotheses, questions, the significance of the study, scope, and limitation of this research.

Section two covers the review of articles, journals, and other print publications from reliable sources dealing with a similar subject matter. This section will discuss the history of AI and the trends behind the investments in this area of research.

Section three outlines various ways businesses can improve productivity by adopting artificial intelligence. Furthermore, discusses an application of machine learning towards achieving this objective.

Section four discusses the methodology used during this research including the research design, characteristics of the research population, the sampling technique used, questionnaire design, the data collection process and limitations of the research conducted.

Section five will present and analyze data collected during the research survey from various respondents.

Section six gives a summary and conclusion to the research.

Chapter 1: Introduction

1.0 Overview

This chapter covers the introduction of this research work. It gives a summary of the objectives, problem, hypotheses, questions, the significance of the study, scope, and limitation of this research. Furthermore, some key terms relating to this research will be defined in this section of the paper.

1.1 Background of Study

Today, technology creates an excitement for the future when we think of driver-less cars, facial recognition, industrials robots, automation, tumor detection, big data and more. These are all problems solved with Artificial Intelligence. It is easier to think of these problems as very specific tasks executed with the use of AI technology. However, putting together all these tasks performed using AI, AI has impacted businesses today enormously and is reshaping the future, one driven by intelligent beings besides humans.

With increasing changes in market trends, globalization, intense competition, and developments in information technology, the business environment are vulnerable to competition and changes. For business organizations, it is essential to improve their performance, profitability, and productivity, while at the same time, meeting the demands of their customers by increasing value of their products and services (Brynjolfsson & Mcafee, 3). In order to sustain competitive advantage, firms are focused on adopting a variety of strategies. The adoption of technology is considered to be an essential and core business strategy of many business organizations in the retail, service, construction and IT sector (Varian, 7). A newly developed technology that is now being used by businesses is artificial intelligence (AI) and machine learning (ML), which emphasizes on automation of the business processes, enhancing operational and functional efficiency, reducing the risks of error, cost optimization and predicting consumer trends and behavior (Varian, 14). Bataller & J. Harris (2) asserts that AI is beneficial for banks, financial institutions, retail sector, construction organizations, and manufacturing industries since it helps in reducing operational costs, increases the productivity of business processes and increases profitability.

The synthesis of literature also suggests that AI has its drawbacks. As asserted by Stephen Hawking in the 2016 issue of Guardian, the increase used of AI in forthcoming time is most likely to increase unemployment rates among lower to middle-income households with only a few roles such as "caring, creative or supervisory roles remaining" (Hawkings, 2016). As asserted by Knapton (2016), middle-class workers will be replaced by machines in forthcoming time. Regardless of its drawbacks, AI is considered to the fundamental driver of the fourth stage of industrial revolution since it is responsible for connecting the physical, virtual and digital spaces. This research aims at analyzing the concept of AI and its significance in the lights of broad and diverse academic resources. Furthermore, it will discuss and demonstrate how it can be adopted towards improving business productivity, as well as highlight its benefits and drawbacks with a specific focus on business.

1.2 Statement of Research Problem:

Artificial Intelligence has gained popularity across various industries and is seen as the future towards sustainability and improvement in businesses. However, many are skeptical about the rapid evolution in AI and deem it to be a threat rather than an opportunity for the industry.

In today's economy, the sustainability of a business can be related to its level of productivity over time. It has become increasingly challenging to maintain consistency in or to improve productivity due to the high demand to meet customer expectations, the rapid evolution of technology and globalization. Thus, the need to harness a technology that offers automation of business processes, insights into customer preferences, reduced cost on resources, increased levels of consistency and more.

This research is set out to educate its readers on current industry practices adopted in the area of Artificial Intelligence, the opportunities and the threats this technology presents to the industry, and an application of Machine Learning, a subset of AI in solving business problems.

1.3 Research questions:

The following questions will help shape this research.

- What is Artificial Intelligence?
- What is Machine Learning?
- Why is Artificial Intelligence relevant in business today?
- Is AI a threat or an opportunity for the industry?
- Can Machine Learning be integrated into various business functions?
- Can Machine Learning solve critical business problems that affect sustainability?
- What is the future of Artificial Intelligence in business management?

1.4 Research Hypotheses:

The following hypotheses are derivatives of the research questions framed.

1. H1: Artificial Intelligence plays a significant role in businesses today and its attributes are related to technological advances, globalization and customer expectations.
2. H2: Machine Learning is a subset of Artificial Intelligence that provides useful techniques through algorithms that can be adopted towards improving business productivity.

1.5 Research objective:

This research is tailored to provide answers to the research problem and research questions.

The first objective of this research will be to explore the history and present-day application of AI in the industry, impact, opportunities, and threats.

The second objective is to design and build a machine learning algorithm that helps businesses improve productivity by automating the customer support process through the use of a ChatBot that can be fully integrated into a business process.

1.6 Significance of the study:

This research is significant because of the enormous growth and pace of evolution in Artificial Intelligence over the past decade. As such, getting acquainted in the various trends, development, and opportunities presented by AI can help businesses redefine, create or optimize processes that lead to increased levels of productivity. This means that's better services, products or solutions can be delivered in an ever-changing industry where the customers are demanding and technology is always evolving. This research will help businesses to think, innovate or reinvent their products, services or solutions.

1.7 Scope/Limitation of the study:

This research is focused on improving business productivity with the use of artificial intelligence, applying research findings to solve similar problems encountered in companies across other various industries. The study conducted is focused on difficulties businesses face in other to keep up the levels of productivity while meeting the customer demands consistently. This study is accomplished by a survey issued to experts across various industries in other to better understand the scope of problems encountered and to define an approach to be adopted for the research. The survey questions were strictly framed to findings extracted from secondary sources. The research spanned a period of three months.

1.8 Definition of key terms:

Artificial intelligence: This is the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.

Machine learning: Machine learning is a field of artificial intelligence that uses statistical techniques to give computer systems the ability to "learn" from data, without being explicitly programmed.

Deep learning: Deep learning is part of a broader family of machine learning methods based on learning data representations, as opposed to task-specific algorithms.

Algorithm: This is a process or set of rules to be followed in calculations or other problem-solving operations, especially by a computer.

Data: These are facts and statistics collected together for reference or analysis.

Information technology: The study or use of systems (especially computers and telecommunications) for storing, retrieving, and sending information.

Business process: This is defined as a set of activities and tasks that, once completed, will accomplish an organizational goal.

Productivity: The effectiveness of productive effort, especially in industry, as measured in terms of the rate of output per unit of input.

Computational linguistics: The the branch of linguistics in which the techniques of computer science are applied to the analysis and synthesis of language and speech.

Robotics: The branch of technology that deals with the design, construction, operation, and application of robots.

Communication infrastructure: The technology, products and network connections that allow for the transmission of communications over large distances.

Knowledge AI: is a form of artificial intelligence (AI) that aims to capture the knowledge of human experts to support decision-making.

Chapter 2: Literature Review

2.0 Overview

This section of the research will focus on the history of Artificial Intelligence, notable achievements of Artificial Intelligence in modern history and the trend behind the massive investments done by companies in this new technology.

2.1 History of Artificial Intelligence

The concept of developing machines that exhibit human-like behaviors dates to 1956 historically at a conference in Dartmouth College. This new concept of having a technology that can replace humans was first introduced at this period with a few notable applications that were based on logic theorems, which were at a time distinguished from the geometric forms used in intelligence tests. An example of this application was demonstrated with a chess game aided by AI. This led to the idea that intelligent computers that could think and make human-like decisions could be created. Between the years 1950 to 1960, scientists from various fields began to discuss the possibility of creating an artificial brain. Also, the field of artificial intelligence research was founded as an academic discipline at this time.

In the 1950s, Alan Turing introduced the concept of the Turing Test. A Turing machine is an abstract machine that was designed to construct any algorithmic logic. Due to the early discoveries in neurology, information theory, and cybernetics, the idea of creating an electronic brain was coined by Alan Turing. This was deemed the first major proposal in the field of artificial intelligence.

Following this proposal and the growing interest in the field by various scientists, the laws of robotics were proposed by Isaac Asimov stating as follows; a robot may not harm a human being, a robot must protect its own existence and a robot must obey the orders given to it by humans, except such orders conflict with the first law (Copeland).

In the following years, the first self-learning game program was created, the logic program for solving mathematical programs was introduced. By 1958, John McCarty from MIT created the List Processing Language and by 1959 the first MIT lab dedicated to artificial intelligence was set up.

After a decade, the first robot was introduced by General Motors (GM) assembly line. The aim of this robotic arm was to aid factory automation. The robot was called Unimate 001 and by 1961, was the first mass-produced robotic arm for factory automation. This was an invention of George Devol and sold as a product of the Unimation incorporation (Agarwal).

In 1964, the industry saw the first demonstration of an AI program that understands Natural Language. Following this was the invention of the first Chabot, an artificial intelligence program ELIZA. At this time, there is a considerable amount of research ongoing areas that were needed to improve the decision-making process and rationale. By the end of 1978, Herbert Simon earns a Nobel prize for his contribution towards the rationality theory that was an important work of artificial intelligence (Anyoha).

The years between 1980 and 1999 saw developments personal computing devices by IBM, the first autonomous vehicle using neural network by Carnegie Mellon, the production of a human­looking robot “Cog” by MIT in 1993 and early introductions of emotional AI in 1999.

In the modern era “2000's”, there has been significant progress in artificial intelligence with the introduction of gestures and human-like movements in robots, self-driving cars, autonomous vehicles and the exploration of emotional intelligence.

Looking back at history, AI has been around for quite a while but is evolving at a rapid pace due to other technological advancements such as the introduction of faster computing power, the exponential growth in digital data, the improvements in communication infrastructure and advancement in AI research.

2.2 Major achievements of Artificial Intelligence in modern history

AlphaGo

This is considered the most complicated game ever invented. It offers an incredible number of possible actions and relies on human intuition. Google's Deep Mind AlphaGo taught itself the game by playing millions of parties with their copies and in 2016 was able to defeat the Go champion Lee Sedalia in four games out of five (Schultebraucks).

Tesla driver-less car

Tesla is a popular name and brand in the industry today. One of the core technological factors that drive its innovation is AI. Tesla CEO Elon Musk stresses that cars with autopilot system are safer than cars without it. A report from the National Security Council in the United States indicates that in the year 2015, the number of traffic fatalities was 1.3 cases for every 100 million miles traveled by conventional vehicles, while the indicators of Tesla - 130 million miles had only one accident. At the same time, the autopilot of the company had saved a life, in a scenario were the car delivered his master Joshua Nellie to a hospital when he suffered a sudden heart attack while on the road (Anyoha).

Siri language

Today it is one of the most high-profile fruits of machine learning, along with similar products from Google (the Assistant), Microsoft (Cortana), and Amazon's Alexa. This technology has changed the way we interact with our devices in a way that would have seemed impossible a decade ago (O'Malley).

AI for event prediction

Humans are fascinated by the concept of predicting a future event given current and historical data. AI has been heavily involved in making this a near reality through the analysis of the massive amount of data, and the identification of patterns or relationships that give useful insights on the probability of an event occurring. An example of this is the case of an Indian company MogIA that embarked on a project to predict the next American president in 2016. The company used an AI tool to analyze over 20 million social media data and was able to further determine to a high degree the voters' support and trust towards a specific candidate.

The results of the election came as a surprise to many, including insiders of the political system. However, the Indian startup MogIA confidently predicted a victory for Donald Trump.

Artificial intelligence has revolutionized the diagnosis of cancer

Health care is one of the areas in which the successes of AI has the greatest practical value. AI has been adopted in a rather very practical means to study illnesses, patients, establish possible relationships between the causes, symptoms, treatment, and the patient. A case is seen were IBM Watson was already able to see deviations in the health of an individual, which elude the attention of experienced diagnosticians. Statistically, about 30 % of cases were Watson puts patients with an additional diagnosis are missed by medical professionals (Schultebraucks).

2.3 Investments in AI for the future

2.3.1 Robotic workforce

It is no more a secret that much of the labor-intensive work in assembly lines of factories would be done by AI programmed robots and not workers in the future. This would bring down the cost of hiring workers and reduce outsourcing. There have been some notable investments in this area of AI with recently, a Chinese T-shirt manufacturing company signing a memorandum of understanding to employ 400 workers in a factory that will use sewing robots to manufacture apparels.

This, however, has been slowly gaining popularity in Japan where a large percentage of elderly care is designed to be executed by robots by the year 2025 (Howal).

2.3.2 Ubiquitous Artificial Intelligence

Presently, AI impacts multiple fields and aspects of our daily activities. Companies are pushing to achieve the best in service value, products and solutions by harnessing AI. In the UK, IntelligentX aims to introduce the world's first AI brewed beer. In Russia, DeepFish is using neural networks to identify fish species. In Sweden, Hoofstep is raising funds from venture capitalists in other to introduce deep learning-based behavioral analysis for horses (Chaturvedi).

2.3.3 Voice Assistants

Voice-enabled computing has been adopted largely in the electronics industry. Most of the IoT devices today have Amazon Echo or Google Home integrated into it. Samsung is also working on its own voice assistant, Bixby. The aim is to have all of its device internet-connected and enabled with the intelligence from Bixby by 2020.

2.3.4 Capsule Networks

One of the most popular neural networks is known as convolutional neural networks(CNN). The CNN is a class of deep learning and is commonly applied to analyze visual imagery using a variation of multiple layer perceptron. Now a new architecture, capsule networks (CapsNet), has been developed and it would outpace the CNNs on multiple fronts. CNNs have certain limitations that lead to a lack of performance or gaps in security. Capsule Networks would allow AIs to identify general patterns with fewer data and be less susceptible to false results. Capsule Networks would take relative positions and orientation of an object into consideration without needing to be trained exhaustively on variations (Zhang).

2.3.5 AI medical diagnostics

Regulators in the US are looking forward to approving AI for use in clinical situations. The advantage of AI in diagnostics is early detection and better accuracy. Machine learning algorithms can compare a medical image with those of millions of other patients, picking up on nuances that a human eye may otherwise miss. Recently, an Anglo-Swedish pharmaceutical company, declared a partnership with Alibaba subsidiary to develop AI-assisted screening and diagnostics applications in China. GE and Nvidia have also joined hands to bring deep learning capabilities to GE's medical imaging devices.

2.3.6 Robotic Process Automation

The main aim of robotic process automation is to use scripts in other to automate human action, that will support business processes. Currently, it is used where it's too expensive or inefficient for humans to execute a process. In RPA, developers use business process documentation to map out and build the solution. Process flows are created, and the tool is taught about the target application it needs to interact with. The end-to-end automated business process model is successfully modeled in the system as it easily navigates through the target applications, executing business process just as a human would. Implementation of RPA improves delivery, reduces processing cost and ensures consistent quality.

2.3.7 Text Analytics and NLP

Natural language processing (NLP) uses and supports text analytics by facilitating the understanding of sentence structure, the meaning, sentiment, and intent through statistical and machine learning methods. Currently used in fraud detection and security, there is a wide range of automated assistants and applications for mining unstructured data. Text analytics and NLP is used across various industries like for improving customer care. It is used to provide a rapid, automated response to the customer (chatbots), dramatically reducing their reliance on call center operators to resolve problems. Spam is a major issue for internet service providers as it may be an entry point for viruses and it impacts productivity in various ways. Text analytics can be implemented to improve the effectiveness of statistically based spam filtering methods.

2.3.8 Natural Language Generation

Natural Language Generation (NLG) is focused on producing text from computer data. It acts as a translator and converts the computerized data into a natural language representation. In this, a conclusion or text is generated based on the collected data and input provided by the user. An example of an interactive use of natural language generation is the WYSIWYM framework, which stands for “What you see is what you meant”. It allows users to see and manipulate the continuously rendered view (NLG output) of an underlying formal language document (NLG input), thereby edit the formal language without learning it. Currently used in customer service, report generation, and summarizing business intelligence insights. The most successful applications have been content generation systems that assist human writers and makes the writing process more efficient and effective.

2.3.9 Speech Recognition

Speech recognition is used to transcribe and transform human speech into a format useful for computer applications. Currently used in interactive voice response systems and mobile applications. Remarkable applications of speech recognition can be seen in Amazon Echo, and Google's Home Assistant. Another Google's speech-recognition product is the AI-driven Cloud Speech-to-Text tool which enables developers to convert audio to text through deep learning neural network algorithms. Working in 120 languages, the tool enables voice command-and-control, transcribe audio from call centers and to process real-time streaming or pre-recorded audio.

2.3.10 Decision Management

Engines that insert rules and logic into AI systems and are used for initial setup/training and ongoing maintenance. A mature technology, it is used in a wide variety of enterprise applications, assisting in or performing and executing automated decision-making processes. These systems help in bridging complex and dynamic information to generate hypotheses. In the finance industry, decision management systems with AI also act as a fraud detector. AI can play a defining role in decision management systems by automating the requirements of the financial markets, pushing the current limits, and taking the competition to the next level. Also, some medical companies are collaborating their decision management systems with AI for providing recommendations for the treatment of specific patients.

2.3.11 Deep Learning Platforms

A special type of machine learning consisting of artificial neural networks with multiple abstraction layers. Currently is primarily used in pattern recognition and classification applications supported by very large data sets. Amazon Web Services (AWS), Google Cloud Platform and Microsoft Azure have created several machines and deep learning API's and microservices that will make it easy for businesses to deploy AI for business operations and automation purposes. These solutions will have the same advantages as the vendors' other service offerings; they will be cost-effective, easy to set up and quick to deploy, making them attractive options for companies that do not have highly skilled, in-house developers.

2.3.12 The arrival of “backprop”

Backpropagation is the single most important algorithm in the history of machine learning. What backpropagation does is to allow a neural network to adjust its hidden layers if the rendered output doesn't match the expected result. In short, it means that creators can train their networks to perform better by correcting them when they make mistakes. When this is done, backprop modifies the different connections in the neural network to make sure it gets the answer right the next time it faces the same problem (Forbes Magazine).

Chapter 3: Improving Business Productivity with AI

3.0 Overview

This chapter discusses ways AI can be harnessed to solve various challenges in business today. In this section, there is a demonstration of a solution that attempts to solve problems in businesses. This solution is developed with the help of existing machine learning algorithms and can be integrated as a business process or used as a tool for better service delivery.

3.1 Introduction

The analysis of literature suggests that there is no standard definition of AI. Many of the researchers have given a generalized definition for it. The generalized definition of AI is a set of information technology and computing based technologies, methodologies, and approaches that make use of the computer and IT devices to reach to flexible and logical decisions in order to respond to environmental conditions that are unpredictable in nature. According to Tredinnick (38), AI refers to IT systems that have the ability to perceive, feel, understand, act and acquire knowledge. AI uses a variety of technologies that permit computers to sense the physical world, evaluate it and collect information, make rational and logical decisions and make recommendations and facilitate learning. A variety of technologies associated with AI have been identified (Tredinnick, 39). These include sensor procession, audio procession, natural language procession, knowledge representation, expert systems, inference systems, and machine learning. Many of the AI technologies are combined together for better output (Skilton & Hovsepian,123). For instance, search algorithms used with machine learning are used to reduce classification errors, while at the same time, increase optimization for parameters. Machine learning aid search algorithms to understand and enhance their fundamental heuristics (Skilton & Hovsepian,123). A number of surveys and reports have revealed the potential of using AI in the field of business. According to the report compiled by Organization for Economic Cooperation and Development (OECD) in 2016, it is expected that AI would be used extensively by business organizations to automate their processes, improve their decision making and boost their performance in the next two decades. A report compiled by MIT Sloan Management Review and Boston Consulting Group revealed that many of the businesses who have understood the potential of AI have implemented it within their organization to address the organizational challenges, improve business performance and enhance business processes (Columbus, 2017). Key trends identified in the report are summarized as follows:

- It is expected that organizations that use AI can successfully Increase in product diversity and offerings in the next 5 years as compared to firms would not use AI (Columbus, 2017).
- AI is expected to improve advertisement and marketing related activities through market automation. Furthermore, it will be beneficial in improving the service provided and supply chain efficiency (Columbus, 2017).
- AI adopted at its initial stage can help organizations to sustain competitive advantage in the long run (Columbus, 2017).

According to Jia et.al (3), the Chinese government had been successful in identifying the potential benefits of AI in the field of business and therefore, has used implemented as part of a governmental strategy to develop technological capabilities of businesses in China. Chinese companies such as Baidu, Ali Baba and Tencent are using AI as part of their business strategy to improve the efficiency of their online services and overall productivity. According to Jia et.al (5), Baidu has adopted AI technology since 2013 in China and launched two core business initiatives: Baidu Cloud and Baidu Brain. The former acts as infrastructure to collect, analyze and tag data. The latter is responsible for a platform for an algorithm. Ali Baba has adopted AI to improve the efficiency of its e-commerce channels and payment systems (Jia et.al, 6). Tencent uses AI technology to manage its social network data through speech and image recognition (Jia et.al, 6). Furthermore, it uses WeChat to provide support to customers (Jia et.al, 6).

3.2 Adopting AI in Business: Benefits

The literature review suggests that the use of AI offers many benefits to a business. They are discussed as follows:

3.2.1 Data Prediction and Analysis

Literature suggests that the use of AI methods such as machine learning (ML) is applicable to data for data prediction (Deb, Jain & Deb, 2). According to Skilton & Hovsepian (156), the advertisement function of organizations frequently faces several issues in terms of selecting the advertisement that is compatible with the product. From a marketing perspective, it is essential that the product's characteristics are directly aligned with the consumer desires and requirements and therefore, marketing aims at making the connection between the two (Pandey, 129). Marketing also emphasizes analyzing and predicting consumer behavior, which can change with time. This aspect of the market analysis is needed to make informed marketing and advertising decisions to meet short term and long term marketing goals (Pandey, 129). AI technology can be instrumental in analyzing marketing data and creating models based on marketing data to predict consumer purchase behavior and market trends. AI-based technology is further used by the marketing division of the company to connect with the users. For instance, Amazon Echo has been designed by Amazon that used speech and text recognition, allowing users to make purchases through verbal commands (Oana, Cosmin & Valentin,357).

3.2.2 AI and Customer Interactions

Literature has revealed that AI is also used to improve customer interaction and management of customer relationships. Many businesses have adopted AI to improve consumer interactions through automation since previous customer interaction methods were dependent on human workers. Automation through AI can help businesses to automate the consumer communication process through data analysis (Oana, Cosmin & Valentin, 359). Consequently, computers can be programmed to respond to consumers in an effective manner. When AI is combined with machine learning, the consumer-computer communication process improves significantly. For instance, Chatbots can be used effectively to communicate with several consumers at the same time. Cangemi &Taylor (2) assert that by 2020, 80% of the consumer interactions will be managed by AI machines.

3.2.3 AI and Business Process Automation Benefits

Cangemi &Taylor, (3) assert that companies can successfully use AI to improve profitability and sustainability of their businesses through the reduction of operational costs and increasing value. AI is responsible for improving efficiency and effectiveness of business processes since it automates the process and reduced the probability of error (Norman, 26). Business automation through AI helps in automating processes and tasks and aligning them with business goals, which ultimately helps in improving operational efficiencies and productivity (Norman, 27). Furthermore, setting up intelligent systems help in reducing human error. The vast amount of data is managed efficiently by these systems, allowing the organization to make better decisions.

3.3 Disadvantages of AI in Business

Although AI offers significant benefits for businesses to grow, there are several disadvantages associated with it. These are discussed as follows:

3.3.1 Cost

The implementation of AI in a business organization is the cost associated with it. The use of smart technologies such as AI is considered to be extremely expensive since they are complex in nature and require competent and trained staff for its maintenance and repair (Klumpp, 224). AI related software programs require a regular update and are vulnerable to changes because of changing business environment (Klumpp, 225). Consequently, updating the system frequently can lead to the risk of losing valuable data. Restoration of the system has been identified as a costly and time­consuming process.

3.3.2 Technical Challenges

According to Klumpp (226), the implementation of AI systems within the workplace is considered to be a lengthy process. Furthermore, these systems are complex in nature and therefore, require technical experts that can manage them more effectively. Konish(26) asserts that integration challenges often come within the workplace related to the implementation of these systems because of lack of understanding.

3.3.3 Privacy Issues

The implementation of AI systems within the business organization comes with issues of privacy and transparency (Klumpp, 224. AI software, like other IT software, is vulnerable to attacks. Consequently, private data can be accessed and misused by third parties. AI systems, therefore, require strong and robust privacy measures that would emphasize protecting user data (Konishi,26- 28).

3.3.4 Unemployment and AI

A series of ethical issues pertaining to AI have been identified with respect to business. For instance, automation can lead to increasing unemployment rates in the forthcoming time since machines will replace human workforce (Knapston, 2016).

3.4 An Application of Machine Learning in Business

Machine learning methods are frequently used in business organizations to process data in order to come up with a prediction. The use of different machine algorithms has been emphasized in the literature. It is very useful in business since it can effectively identify patterns that cannot be identified by human beings (Lichtenthaler, 13).

3.4.1 ChatBot

A ChatBot is considered to be machine learning and artificial intelligence based program that has the ability to make conversations with humans through speech recognition or text recognition technology (Lichtenthaler, 14). These programs have been designed and taught to behave like human beings and therefore, used by businesses to communicate with their potential consumers. Chatbots are generally based on text recognition technology, which is responsible for acquiring information from the consumers as well as provides customer services by responding to their queries. Literature suggests that some of the ChatBots are based on natural language processing systems (Lu et.al, 370). However, simple ChatBots systems are also available. Google Assistant, Apple Siri, and Amazon Alexa are some of the ChatBots that are being used by organizations to communicate directly with the consumers (Lu et.al, 372).

3.4.2 Improving the business process with an AI ChatBot

Chatbot has been used extensively by several international companies to improve their interactions with their consumers. For instance, chatbots at Pizza Hut via Facebook messenger allows their customers to place orders (Bataller & J. Harris. 8). These chatbots collect information such as delivery areas and dietary information. Amazon application programming interfaces are being used by FedEx to tell consumers regarding their services and offerings (Brynjolfsson & Mcafee, 10). Amazon Echo device is being used by hotels to connect with consumers by providing them with information related to rooms, hotels, recreational activities, and other services. H&M chatbot KiK offers customer service, product information, style tips, and clothing information to its consumers (Brynjolfsson & Mcafee, 11). Furthermore, the chatbot also allows consumers to make transactions to make online purchases. In the hospitality sector, chatbot of Dutch airline KLM uses Facebook Messenger to confirm a booking and offers flight information services to its customers (Cangemi &Taylor, 5). Taco bell's chatbot TacoBot provides customers to read the menu and specify the order pickup from any location (Varian, 32).

3.4.3 Customer Demands

Since consumers require fast customer service, staffing live agents can be a hassle and problem since even for larger organizations, the agents cannot answer queries and tickets in a timely manner (Varian, 87). Therefore, chatbots are frequently used to engage directly with consumers to provide the product or service information. Furthermore, they are used to respond to consumer queries. Chatbots is a useful tool that improves consumer response rate at 90% rate (Mehrotra, 357). This is because they have the ability to chat with several consumers at the same time. According to Mehrotra (359), chatbots are instrumental in converting visitors into consumers because of their efficient and high response rate.

3.4.4 Drawbacks

Chatbots are expensive and cannot fully replace human live agents. Their implementation requires technical expertise. According to Tredinnick (40), chatbots need to be trained to understand the consumer requirements, which is the fundamental problem in their design. Different organizations have different consumer segments and therefore, chatbots need to be trained accordingly, otherwise, it can affect the business in a negative manner (Lu et.al, 372).

3.5 AI ChatBot Design and Implementation

3.5.1 Business Requirement

Design and implement a solution that will minimize the waiting time required for customers, clients and employees to get information technology support on the various products and solutions in operation within the company. This solution should help optimize the IT support process and as well provide the basis for obtaining customer feedback on the services periodically.

3.5.2 System Architecture

The conceptual model that will help define the structural design and behavior of the system is illustrated below. The diagram below illustrates the various components within the system and how the interaction between components are handled.

3.5.3 Conversation Process

The conversation process adopted for the implementation of the ChatBot solution is illustrated visually below. This process requires the user to initiate a conversation with the AI bot through input methods such as text or voice commands. The system will extract the keywords and phrases based on this input, the intent and event of the extracted phrases will be analyzed and a suitable response will be shown (output) to the user based on the machine learning algorithm and training given to the agent. This process can be iterative based on the satisfaction of the user and can be terminated at any given time by the user.

3.5.4 Dataflow Diagram

This section illustrates the flow of data within the system. The platform adopted for the implementation of this project is a Google-owned technology for developing human-computer interaction (HCI), that is based on natural language conversations. The platform Dialog flow provides the basis for developers to design and implement conversational artificial intelligence bots that can be embedded in external applications run on the web or mobile platforms. The diagram illustrates the data flow level zero (context diagram) and the level one of the system.

3.5.5 Conversation Flow

The conversation flow diagram is derived from the business narrative and requirements analyzed. This is a representation of an ideal scenario of a conversation between a human and the AI bot. It takes into account various alternative patterns a conversation may take, with processes in place to end a conversation within a timely manner while meeting the set objective and improving the learning process. This flow has undergone various iterations and some processes that may not be depicted in entirety at this stage. However, the conversation flow has been tested logically and can be implemented with set goals, training methods and continuous improvement as the project implementation scales.

3.5.6 ChatBot Prototype 1 Demonstration

The demo version of the ChatBot application is accessible at this web address www.mbailcidemochatbot.kubixtech.com . This prototype demonstrates an implementation according to the design and process presented in this chapter. This system can support hundreds of users accessing the support desk at the same time. With the help of fast, secure and reliable infrastructures in place, the technical support unit can adopt this tool to satisfy the business requirement presented in the subsection 3.5.1 of this chapter. However, for this system to meet user expectations, it required extensive training and continuous improvements in the logic and rules that guide its decision-making process.

Abbildung in dieser Leseprobe nicht enthalten

Chapter 4: Research Methodology

4.0 Overview

This chapter discusses the methodology used during this research. The following techniques were used as an adopted methodology to accomplish this research: a review of literature in relation to artificial intelligence, questionnaire for gathering data, interviews with some professionals within various industry, data analysis of data gathered and the development of a prototype that emulates a problem and simulates a solution that solves the problem. This chapter will discuss the research design, characteristics of the research population, the sampling technique used, questionnaire design, the data collection process, and limitations of the research conducted.

4.1 Research Design

This research consisted of 5 stages:

- Phase 1- At this stage, the problems were defined, and objectives were established in order to develop a research plan that would be used as a baseline. At the end of this stage, a topic for this research was proposed, a research plan was developed and the research objectives were clearly established along with research problems.
- Phase 2 - At this stage, a literature review was conducted in relation to artificial intelligence and ways of improving business productivity with an application of machine learning.
- Phase 3 - This stage of the research included the questionnaire design, field Survey and a technical design and implementation of a chatbot prototype.
- Phase 4 - This stage of the research included the distribution of questionnaires in other to gather required data to meet research objectives.
- Phase 5 - This stage focused on data analysis, discussion, conclusion, and recommendations. [Due to copyright reasons this image has been removed by the editorial staff.]

4.2 Research Period

This study began in May 2018 when the initial proposal and topic was approved. The literature review was completed by end of July 2018.

The research review, questionnaire design, and strategy, chatbot design and prototype implementation were completed in the first week of October 2018.

The research distribution, analysis of data gathered, discussion and conclusion spanned a period of five weeks and was completed in the second week of November 2018.

4.3 Characteristics of the Research Population

The research population included IT managers, various IT professionals both in research and development fields of expertise, and individuals with experience in IT related projects. Considering the targeted population, participants in this survey were selected randomly.

The targeted population consisted of individuals and groups, who are involved in research or development projects relating to artificial intelligence. These individuals and experts are from diverse geographical locations and organizations and possess experiences through their professional careers in various consulting firms, international agencies, institutions, etc.

4.4 Survey Sampling Technique

In this research work, the study was a descriptive survey and key standard frameworks enabled it to utilize questionnaires among the selected respondents in the field. However, the respondents were diversely classified based on their relevance to the study.

In this method also, there was an important need to also acquire relevant field data from respondents by using interview schedules. These were distributed to individuals whose focus were based on their relevant experiences and input they demonstrated through additional provisions. The study design was aimed at collecting data at a point and using it to describe the nature of the existing condition.

4.5 Data Collection

Questionnaires are widely used today as a medium of data collection for conducting surveys. Questionnaires are used for descriptive and analytical surveys to discover facts and various opinions. This is the basis for which a questionnaire was chosen as a technique to be used for collecting data during this research.

During the research, data is collected in a standardized form from the sample population. This allows the researcher to carry out statistical analysis of data collected via various tools, for example, computer software analytical tools.

4.6 Limitation of Research

This research was carried out about the major objectives of the study. However, as established by the field study, key challenges were encountered, including poor coordination of the research and inadequate manpower to carry out the research. Furthermore, the questionnaire contained simple questions with no control over respondents’ answers. In some cases, respondents decided to give a general or nonspecific response to a question.

Chapter 5: Data Presentation, Analysis, and Interpretation

5.0 Overview

This chapter will present and analyze data collected during the research survey from various respondents. The result of the questions included in the survey will be discussed in two sections. Section one will present information on the respondent profiles. Section two will identify the factors that affect stakeholder management during an IT project. This chapter will conclude with a summary of major findings deduced from the survey questionnaire.

5.1 Analyses and Presentation of Primary Data

There were 25 respondents that participated in this survey. However, there were 5 incomplete surveys. Based on the 20 participants with a completed survey, data received was analyzed and presented below according to the survey questions.

The aim of this survey was to analyze the impact of the current support process in the business from the users' point of view.

Research question: ‘Please specify if you are an employee, client/customer, subcontractor, other?'

Results collected from the survey showed that 60% of the respondents were employees of the company, 15% were clients or customers, 25% were subcontractors working with various service providers. A further analysis of the employees shows that 50% were working in the customer service unit, 8% in the supply chain, 8% in operations, 16% in sales and distribution and 16% in marketing.

Abbildung in dieser Leseprobe nicht enthalten

Research question: ‘What is the average waiting time when connecting to a support agent?’

The results from the survey shows that 5% of the respondents say it takes a minute on average, 30% say it takes five minutes on average, 15% say it takes less than fifteen minutes on average, while 50% of the respondents are of the opinion that it takes more than fifteen minutes on average.

Abbildung in dieser Leseprobe nicht enthalten

Research question: ‘How long does it take on average to resolve an issue you have reported?' Based on the results collected, 5% of the respondents say it takes less than fifteen minutes, 10% of the respondents say it takes less than thirty minutes, 15% say it takes less than an hour, 60% say it takes more than an hour, 10% say it takes up to a day and none for up to a week.

Abbildung in dieser Leseprobe nicht enthalten

Research question: ‘How often do you contact the support desk?'

The results show that 10% of the respondents contact the support desk once per day, 20% once per week, 5% multiple times a day, 45% multiple times a week, 5% every day, 15% during peak season activities.

Abbildung in dieser Leseprobe nicht enthalten

Research question: ‘What channels of communication do you use?’

All respondents use email, 25% make phone calls, 15% use instant messaging and All do use Quick Support by TeamViewer.

Abbildung in dieser Leseprobe nicht enthalten

Research question: ‘What is your preferred channel of communication with the support team?' 75% of the respondents prefer Quick Support solution, none prefer instant messaging, 20% prefer to make direct phone calls and 5% prefer emails.

Abbildung in dieser Leseprobe nicht enthalten

Research question: ‘What is the reason for the preferred option?' 90% of the respondents indicated quick response as a reason while 10% indicated convenience as the reason.

Abbildung in dieser Leseprobe nicht enthalten

Research question: ‘In most cases, what do you contact support for?' 25% indicated assistance, 65% indicated an issue or bug, 10% indicated suggestion.

Abbildung in dieser Leseprobe nicht enthalten

Research question: ‘Rate your experience with our support team' 5% of the respondents say it is poor, 40% say it is okay, 45% say it is good and 10% say it is awesome.

Abbildung in dieser Leseprobe nicht enthalten

Research question: ‘What can be improved within our support process to better serve your needs?' 70% of the respondents think a timely support would improve the process while 30% suggest that more human resources be hired.

Abbildung in dieser Leseprobe nicht enthalten

5.2 Summary of findings

In conclusion, the following findings can be deduced from the survey questionnaire;

1. There was a mix of technical and non-technical respondents and the response received varied depending on the various job roles and experience of the respondents.
2. Based on the survey, a majority of the respondents shared the view that the current support process is time demanding and often may require more frequent contacts to be made daily.
3. It can also be deduced that it takes a long time for users to get issues resolved.
4. Furthermore, 45% of the respondents are not satisfied with the current level of support rendered.
5. Most of the respondents’ demand quicker response time while a good number suggest that more resources should be allocated to the technical support team.

Chapter 6: Conclusion

6.0 Conclusion

Artificial intelligence is being used extensively in businesses to improve business performance, profitability, and productivity. Businesses need to survive in the market and therefore, in the era of intense competition, globalization and changing marketing conditions, it is essential for firms to adopt technologies such as AI to succeed in the market. A variety of organizations that operate in different sectors such as retail, services, IT and construction use AI as part of their organizational strategy to improve their business operations and functions, reduce errors, reduce operational costs and to analyze consumer trends. AI has undoubtedly a lot of potentials in business. It can increase product portfolio, improve advertisement and marketing related activities through automation and improve the overall level of service delivery. AI can extensively be used to predict data. Organizations can use AI models to determine patterns and analyze models to identify different trends. In an advertisement, automation through AI can help in predicting consumer behavior, consumer purchase intention, and market trends. Consequently, it can improve the firm's decision­making process in terms of new product development. AI improves consumer interactions and helps in customer relationships. Automation can help in dealing with consumer queries effectively. Although AI offers several benefits, it also has drawbacks. It is expensive in terms of its implementation and maintenance. AI requires superior technical expertise and is prone to change with changing technological and business trends. These systems are complex in nature and are vulnerable to attacks, which can lead to privacy issues and data access by third parties.

A ChatBot is a type of AI program, which has been discussed and highlighted in this research. It is frequently used by businesses to communicate with consumers in order to respond to their queries and to provide the product or service information. Chatbots are now being used worldwide since they have the ability to chat with several consumers at the same time. Chatbots are expensive. Their design requires extensive logical precision and training according to different consumer requirements.

6.1 Recommendations and Suggestion for further research

In the field of AI, recommendations for future research are discussed as follows:

Future research should focus on the needs to improve AI technology in terms of its implementation and maintenance since it is extremely expensive. Economical AI systems should be designed so that small and medium businesses can reap the benefits of AI.

Future research should also emphasize on investigating the link between AI and unemployment. This is considered to be a serious problem that can lead to political and socio-economic issues in the long run.

Bibliography

Agarwal, Rachit. 10 Examples of Artificial Intelligence You Are Using in Daily Life. 21 September 2018. 29 September 2018 <www.beebom.com/examples-of-artificial- intelligence>.

Anyoha, Rockwell. The History of Artificial Intelligence. 28 August 2017. 1 September 2018 <sitn.hms.harvard.edu/flash/2017/history-artificial-intelligence/.>.

Chaturvedi, Aditya. 13 Major Artificial Intelligence Trends to Watch for in 2018. 10 July 2018. 1 August 2018 <www.geospatialworld.net/blogs/13-artificial-intelligence-trends- 2018/>.

Copeland, B. J. Alan Turing. 23 August 2018. 1 October 2018 <https://www.britannica.com/biography/Alan-Turing>.

Forbes Magazine. Top 10 Hot Artificial Intelligence (AI) Technologies. 29 March 2017. 1 September 2018 <www.forbes.com/sites/gilpress/2017/01/23/top-10-hot-artificial- intelligence-ai-technologies/>.

Howal, Sagar. AI Writes | The History of Artificial Intelligence - Towards Data Science.” Towards Data Science, Towards Data Science. 4 January 2018. 8 July 2018 <towardsdatascience.com/ai-writes-the-history-of-artificial-intelligence-4d585b537498>.

O'Malley, James. The 10 Most Important Breakthroughs in Artificial Intelligence. 10 January 2018. 23 August 2018 <www.techradar.com/news/the-10-most-important-breakthroughs-in- artificial-intelligence>.

Schultebraucks, Lasse. A Short History of Artificial Intelligence. 17 December 2017. 1 June 2018 <dev.to/lschultebraucks/a-short-history-of-artificial-intelligence-7hm>.

Zhang, Leona. "5 Major AI Trends of 2018 - DZone AI." 12 June 2018. DZone AI. 1 September 2018 <www.dzone.com/articles/5-major-ai-trends-of-2018>.

Bataller, C., and J. Harris. "Turning artificial intelligence into business value."Today: Accenture Emerging Technology Group (2018).

Brynjolfsson, Erik, and Andrew Mcafee. "The business of Artificial Intelligence: What it can--and can not--do for your organization."Harvard Business Review Digital Articles 7 (2017): 3-11.

Cangemi, Michael P., and Patrick Taylor. "Harnessing Artificial Intelligence to Deliver Real­Time Intelligence and Business Process Improvements."EDPACS 57.4 (2018): 1-6.

Columbus, Louis. “How Artificial Intelligence is Revolutionizing Businesses in 2017”. Forbes Website, 2017 https://www.forbes.com/sites/louiscolumbus/2017/09/10/how-artificial-intelligence-is- revolutionizing-business-in-2017/#36a55d925463

Deb, Suman Kumar, Ruchi Jain, and Varsha Deb. "Artificial Intelligence—Creating Automated Insights for Customer Relationship Management."2018 8th International Conference on Cloud Computing, Data Science & Engineering (Confluence). IEEE, 2018.

Norman, Don. "Design, business models, and human-technology teamwork: As automation and artificial intelligence technologies develop, we need to think less about human-machine interfaces and more about human-machine teamwork."Research-Technology Management 60.1 (2017): 26-30.

Jia, Kai, et al. "The Application of Artificial Intelligence at Chinese Digital Platform Giants: Baidu, Alibaba, and Tencent." (2018). ETLA Reports No 81:1-14

Hawking, S (2016) This is the most dangerous time for our planet. The Guardian, 1 December 2016. Available at: https://www.theguardian.com/commentisfree/2016/dec/01/stephen-hawking-dangerous- time-planet-inequality

Knapton, S (2016) Robots will take over most jobs within 30 years, experts warn. The Telegraph, 13 February 2016. Available at: http://www.telegraph.co.uk/news/science/science-news/12155808/Robots-will-take-over- most-jobs-within-30-years-experts-warn.html (accessed 10 November 2018).

Klumpp, Matthias. "Automation and artificial intelligence in business logistics systems: human reactions and collaboration requirements."International Journal of Logistics Research and Applications 21.3 (2018): 224-242.

Konishi, Yoko. "Toward Artificial Intelligence Technologies as an Integral Part of Service Business Operations."RIETIhighlight 68 (2018): 26-28

Lichtenthaler, Ulrich. "Substitute or Synthesis: The Interplay between Human and Artificial Intelligence."Research-Technology Management 61.5 (2018): 12-14.

Lu, Huimin, et al. "Brain intelligence: go beyond artificial intelligence."Mobile Networks and Applications 23.2 (2018): 368-375.

Mehrotra, Prakhar. "Applications of Artificial Intelligence in the Realm of Business Intelligence."Utilizing Big Data Paradigms for Business Intelligence (2018): 1.

Oana, Oprisan, Tileagä Cosmin, and Niju Claudiu Valentin. "Artificial Intelligence-A New Field of Computer Science Which Any Business Should Consider."Ovidius University Annals, Economic Sciences Series 17.1 (2017): 356-360.

Pandey, Peeyush Kumar. "Role of Artificial Intelligence in Business."Economic Development of India 1.1 (2018): 122-139.

Skilton, Mark, and Felix Hovsepian. The 4th Industrial Revolution: Responding to the Impact of Artificial Intelligence on Business. Springer, 2017.

Tredinnick, Luke. "Artificial intelligence and professional roles."Business Information Review 34.1 (2017): 37-41.

Varian, Hal. "Artificial intelligence, economics, and industrial organization."The Economics of Artificial Intelligence: An Agenda. University of Chicago Press, 2018.

[...]