An AI Primer for Managers

A White Paper by Dr. Peter Green

Abstract

Artificial Intelligence (AI) has become one of the most widely discussed technologies in modern business, yet it is also one of the most misunderstood. This paper distinguishes between AI systems that automate specific knowledge tasks using well-established techniques such as rules, planners, intelligent agents, and statistical correlators, and attempts to achieve Artificial General Intelligence (AGI) through large generative AI models.

Drawing on more than fifty years of practical experience designing and implementing AI systems for defense, manufacturing, healthcare, and supply chain applications, the author argues that the greatest economic value of AI lies in replacing repetitive, procedural "intelligent grunt work" rather than attempting to substitute human judgment and experience.

The paper explains the strengths and limitations of the principal AI algorithms, examines why generative AI often fails to deliver a positive return on investment for general knowledge tasks, and discusses the fundamental training-set and economic challenges facing AGI. It concludes with practical guidance for managers seeking to deploy AI successfully by focusing on applications that deliver measurable improvements in productivity, operational efficiency, and business performance while avoiding costly attempts to automate complex human judgment..

Key Take-Away Points

1. Distinguish Between Specific Knowledge and General Knowledge Tasks. AI delivers its greatest value by automating Specific Knowledge tasks that follow defined procedures -not General Knowledge tasks that require human judgment, experience, and common sense.
2. Automate Procedures - Not Judgment. Rules, planners, intelligent agents, and other proven AI techniques can eliminate repetitive "intelligent grunt work," allowing people to focus on higher-value decision making.
3. Generative AI Is a Useful Assistant, Not a Manager. Chat GPT and similar systems are excellent for information retrieval, drafting content, and improving personal productivity, but they should support - not replace - experienced professionals.
4. Measure AI by Business Results, Not by Technology. The success of an AI project should be measured by improvements in productivity, quality, compliance, and return on investment - not by the number of AI tokens consumed or the sophistication of the technology.
5. Successful AI Projects Solve Real Business Problems. Organizations achieve the greatest economic benefit when AI is applied to well-defined operational processes that deliver measurable improvements in efficiency, accuracy, and customer service.

Please click here to download the PDF of
An AI Primer for Managers