Partner, Don’t Police: AI in the Business Classroom
Preface
This book is for business educators deciding where AI-supported practice belongs in a course, and what students must still demonstrate independently. It offers a small, inspectable starting point: a fictional case, source records, contrasting responses, flawed feedback, a checked revision and a changed task.
You do not need a paid AI account to use the book. The worked examples and activities include paper or paired routes. An institution may reasonably decide that a particular tool, dataset or activity is unsuitable.
Reading routes
The sixteen chapters follow four stages:
- Foundations (1–3): the educational purpose, how language models and products differ, and the ethics and data gate to complete before using a service.
- Practice (4–9): getting started, seven prompting patterns, managing context, critique, bounded simulation and self-assessment.
- Design (10–13): assessment evidence, a small content transformation, the InnovateCo case and a unit-level design.
- Implementation (14–16): local adaptation, practical constraints and a decision about a first pilot.
For a first classroom trial, read Ethics and Integrity, then the Pilot Pack and Flight Simulator. For assessment redesign, start with Assessment, the Rubric System and the Stress Test.
The appendices also provide the full prompt library, a facilitator’s workshop guide, institutional alignment, terminology and further reading. They are part of the downloadable editions, not web-only extras.
What the examples establish
PDP-2026.1 is an authored fictional teaching pack, not a report of a classroom trial. Its scores illustrate educator calibration against stated criteria. They do not validate the rubric for another institution or demonstrate improved learning.
The main case concerns a disputed performance issue. It teaches evidence handling and a bounded first response, not qualification to conduct a real investigation. Supplier and marketing examples show how the same checks work outside HR. Adapt the outcome and sources before adopting an example in your discipline.
The book distinguishes supported practice, submitted process evidence and independently demonstrated capability. A polished AI-assisted response, a long transcript and an AI-generated mark cannot substitute for the last of these.
Formats and supporting resources
The online edition is the primary, freely accessible edition. The site provides PDF, EPUB and plain-text downloads, plus the source and related resources.
Paperback ISBN: 979-8-2542-3619-1. Readers who want to support the open edition can look for the paperback on Amazon. Availability and the edition supplied should be checked on the listing; the ISBN is not evidence that a print copy contains the latest source revisions.
The GitHub repository contains the manuscript and history. DeepWiki offers an AI-generated view of that repository; check its answers against the source. Any companion chatbot is similarly a navigation aid, not an authoritative interpreter of policy or assessment requirements. Do not submit confidential material to it.
The downloadable llm.txt can support local search or an approved AI workflow. Its length may exceed a product’s usable context, and uploading it does not ensure every passage will be considered.
Scope and currency
This revision was prepared in September 2026. Product capabilities, institutional rules and legal requirements can change; check the relevant current authority before implementation. Australian examples require local adaptation elsewhere. Nothing here replaces institution-specific approval or professional advice about a real case.
Conversation, Not Delegation explains the wider methodology; this book includes the recap needed to work independently. Find the companion and other open books at books.borck.education.
The series trunk is Conversation, Not Delegation for method and The Human Edge for professional judgement. This book is the branch for teaching with AI.