Artificial Intelligence for Business Optimization

Research and Applications

Bhuvan Unhelkar, Tad Gonsalves

• Publisher: CRC Press
• Year: 2021
• Print ISBN: 9780367638368
• eBook ISBN: 9781000409475
• Edition: 1
• Topic: Informatik
• Subtopic: Unternehmensentwicklung

This book explains how AI and Machine Learning can be applied to help businesses solve problems, support critical thinking and ultimately create customer value and increase profit.

By considering business strategies, business process modeling, quality assurance, cybersecurity, governance and big data and focusing on functions, processes, and people's behaviors it helps businesses take a truly holistic approach to business optimization. It contains practical examples that make it easy to understand the concepts and apply them.

It is written for practitioners (consultants, senior executives, decision-makers) dealing with real-life business problems on a daily basis, who are keen to develop systematic strategies for the application of AI/ML/BD technologies to business automation and optimization, as well as researchers who want to explore the industrial applications of AI and higher-level students.

Chapter 1

Artificial intelligence and machine learning

Opportunities for digital business

Artificial Intelligence in the context of business

Artificial intelligence (AI) is precisely that: it is non-real intelligence that is demonstrated by machines (computers) as they learn and mimic human (natural) intelligence. AI helps businesses achieve their goals based on the parameters provided to its algorithms that analyze vast and relevant data. Success with AI requires an understanding of the business. This is one of the crucial differentiators in a strategic approach to business optimization (BO). Analytics in Big Data are important but not without a proper understanding of business.¹

AI is understood here as a combination of systems, processes, algorithms, and techniques to analyze large, complex, and fast-moving data. The purpose of such analytics is to identify trends and patterns that will help extract insights from the data. Decisions based on the insights are as close to human decision-making as possible and are continuously and iteratively improving on the results. Over decades, thinkers like Davenport² have called AI as “the most important general-purpose technology of our era with wide ranging applications.” This discussion narrows the potentially wide-ranging applications of AI to the one that enables businesses to provide “value” to their customer. An optimized business provides this value in the most efficient and effective way using the tools and techniques of AI. As a result, the business becomes agile. Agility is an important characteristic of business in the digital era.³ BO is thus a strategic application of AI and ML using Big Data in order to provide value to customers with agility. BO is not limited to data and processing. BO uses AI to expand to reach the outer edges of the business wherein it “spots” the customers, understands their needs, personalizes the offerings, and continuously enhances the products or services. AI also helps the business handle governance, privacy, security, and compliance requirements. The keywords of importance are systems, processes, algorithms, techniques, analyze, extracting, insights, large, complex, fast-moving data, support, decision-making, value, efficiency, effectiveness, agility, and customer. Each of these words has a meaning that is specific to the discussion on AI and BO. This book is a journey into these concepts of AI and their application to BO.

Artificial intelligence (AI) and machine learning (ML) as enablers of business optimization (BO)

AI brings together a wide range of technologies, databases, algorithms, and devices as potential enablers of BO. AI provides the necessary optimization capabilities. AI on its own, however, is not BO. BO acknowledges the need to strategically apply AI to business functions. Furthermore, AI is presumed to bring about automation. Contemporary digital business strategies are mainly geared towards automation,⁴ which is not the same as optimization. Optimized businesses are, by necessity, digital businesses but the reverse may not be true. This is so because digital businesses may be automated with AI and ML but not necessarily optimized. Optimization goes further than automation by not merely mimicking the existing processes but ensuring business goals with efficiency and effectiveness. Optimization in this discussion is specifically delineated from automation as a separate and dedicated business initiative.

Big Data provides AI and ML with the necessary range, depth, and variety of data that can be analyzed in order to produce excellence in customer value. BO is not an isolated activity or a project in an organization. BO starts strategically, at the board level, with the examination of the entire business, the environment in which it operates, its functional and structural parameters, and its challenges and opportunities. AI is the enabler to revamp the business based on a long-term view of the business and the environment in which it operates. BO covers people, processes, technology, budgets, security, and quality issues for business and its collaborating partners.

BO initiative starts by studying each business objective and function keeping AI capabilities and constraints in mind. As a result, BO minimizes the risks and maximizes the agility of a business. The decision variables, the objective functions, and the corresponding constraints within a business process are carefully quantified in BO keeping technology restraints in mind. The decision variables serve as inputs to the AI optimization algorithms (e.g., evolutionary computation and swarm intelligence), which can optimize even large-scale business problems in a reasonable amount of time.

Subjective elements in BO

While BO capitalizes on the AI algorithms, it also gives due credence to the nonquantifiable aspects of AI. These unquantified concepts are “subjective.” The acknowledgment and insertion of the subjective element in the optimization of business processes are based on human or natural intelligence (NI). NI is an important part of holistic BO, and it is discussed in Chapter 10. Other issues such as lack of sufficient explainability of AI⁵ are discussed in later chapters.

Machines are very good at mimicking the human functions. Gaming, autonomous driving, robotic surgery, and medical diagnostics are all capabilities that machines can handle well. Computers are phenomenally better at undertaking routine tasks, and the more routine they are, the better is the automation. Increasingly complex tasks require correspondingly complex algorithms which are coded by humans.

Machines, however, lack the cognitive capabilities inherent to humans. As Finlay⁶ puts it, “True AI is about much more than just pattern recognition and prediction.” He further questions “Is there some additional (as yet unknown) element required for human-like intelligence and self-awareness which can’t be replicated via computation alone?”⁷ There is no inkling of that question being answered in the near future. Ada Lovelace expresses this succinctly:

The Analytical Engine has no pretensions whatever to originate anything. It can do whatever we know how to order it to perform. It can follow analysis; but it has no power of anticipating any analytical relations or truths. Its province is to assist us to making available what we are already acquainted with.⁸

The “learning” and “problem-solving” capabilities of AI are programmed by developers. Decision-making uses the analytical insights but is still carried out by people. The quality and goodness in decisions are judged by their consequences which may be subjective and beyond the scope of AI. NI (“humanization”) is a positive influence on AI-based decision-making.

Agility in BO

Agility is an important business characteristic. An agile business is a flexible business that can change according to the changing needs of the users and the environment. Automation, optimization, and humanization are considered together in holistic BO to provide agility to business and enhance customer value. The variances and nuances of business functions, their automation, and their optimization occur on an ongoing basis. AI, ML, DS, and Big Data are considered as the technologies that enable optimization.

Data provides the backbone of the systems and processes of a digital business. Analytics support the optimization by extracting value. Data, however, is exploding. The blinding speed, mounting volumes, and vast geographical reach of digital data result in Big Data. Systems, processes, and techniques that make sense of this Big Data have to be continuously evolving. Business and technologies need agility in order to keep pace with Big Data. Agility in design and implementation of analytics enable not only processing of data but also “remembering” the algorithm execution in an iterative and incremental manner. This ability to “learn” from an execution of the algorithm for varied sets of data is ML. ML design, implementation, and testing require agility. Agility during BO occurs in both the business and the solution space.

AI also personalizes and customizes the visualization of analytical insights based on specific needs of the users, their collaborative choice, and their level of interest. The flexibility in viewing the analytics across business boundaries and over many devices is a part of business processes. AI also maintains consistency and currency of visualization across collaborative businesses.

Collaboration in BO

Another important characteristic of an agile digital business is its extension and reach to other supporting businesses. Multiple businesses are able to collaborate with each other in a distributed and federated environment through digital technologies. Big Data technologies enable the generation and assimilation of dynamically changing contents from a variety of internal or external data sources. AI and ML technologies merge contents developed and published by different sources as appropriate to the required analytics. Decision-makers in collaborative enterprises use data and analytics from a variety of sources on the cloud that may not be owned by them.

Granularity in BO

A crucial factor in providing efficiency and effectiveness in AI is the concept of granularity in data analytics.⁹ The AI technologies enable the analytics to drill down to the fines...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Family
• Table of Contents
• Foreword by Andy Lyman
• Preface
• Readers
• Figures
• Acknowledgments
• Authors
• 1 Artificial intelligence and machine learning: Opportunities for digital business
• 2 Data to decisions: Evolving interrelationships
• 3 Digital leadership: Strategies for AI adoption
• 4 Machine learning types: Statistical understanding in the business context
• 5 Dynamicity in learning: Smart selection of learning techniques
• 6 Intelligent business processes with embedded analytics
• 7 Adopting data-driven culture: Leadership and change management for business optimization
• 8 Quality and risks: Assurance and control in BO
• 9 Cybersecurity in BO: Significance and challenges for digital business
• 10 Natural intelligence and social aspects of AI-based decisions
• 11 Investing in the future technology of self-driving vehicles: Case study
• Appendix A: Frameworks and libraries for ML
• Appendix B: Datasets for ML and predictive analytics
• Appendix C: AI and BO research areas
• Index