AI Work Automation Course

Contents

1. How to use this book
2. Chapter 1. Work analysis and automation design

   Breaking work into steps · Triggers, completion conditions and handoffs · Normal and exception paths · Three criteria: repetitiveness, risk and review cost · Breaking risk down · Work for rules and work requiring AI judgment · The delegation ladder · Where to place human review points · Designing a reviewable workload · Translating work requirements into feature combinations · What not to automate · The one-page automation design sheet · Workflow or agent? · Traps in scoring tables · Rubber-stamp review

3. Chapter 2. Knowledge spaces and custom assistants

   Knowledge spaces (projects) · Three layers of instructions · Shared projects and custom assistants · Why instructions and knowledge are kept apart · Five elements of team instructions and writing tips · Turning a work procedure into instructions · Answers that expose uncertainty and conflict · Knowledge capacity and retrieval · Preparing searchable documents · Organizing materials around questions · Designing update cycles · Testing before rollout · Ingestion, retrieval and generation are different stages · Designs that survive product changes · Boundaries between memory and knowledge

4. Chapter 3. Connectors and agent features

   What connectors and app connections do · Permission inheritance and shared credentials · Authentication, access permissions and execution approval · MCP: the USB-C of AI apps · What to check before connecting MCP · Connecting with least privilege · Tool permission settings: when to ask and what is accessible · Agent mode: plan, act, observe, stop · Agent task specifications · Approval points and the information approval requests must show · Scheduled tasks · Browser-agent risks · Partial failure and duplicate execution · Why indirect prompt injection cannot be completely prevented · Before trusting “Completed”

5. Chapter 4. Sharing, permissions and data boundaries

   Reading permissions across three axes · Sharing permissions by role · Distinguishing sharing units · Group sharing and permission calculation · Sharing pitfalls · Designing sharing scope · Workspace admin controls · Retention, deletion, training use and memory · Data handling by plan · Separating sensitive materials: send less · Derived information and personal versus organizational spaces · Access cleanup for onboarding, transfers and offboarding · Settings usually affect future behavior · Testing whether permissions actually apply · Work automation from a personal-information perspective

6. Chapter 5. Operations and improvement

   Review checklists: choosing items by risk · Error severity and full versus sample review · Aligning reviewers' judgments · Four operational metrics · Metric definitions and denominators · Choosing an evaluation method · Classifying failures by stage · Turning one failure into an improvement experiment · Change management · Measuring benefits · User onboarding and support · Incident response, manual fallback and retirement · Comparison traps · Questioning attractive numbers · Retirement is part of design

7. Appendix A. Glossary
8. Appendix B. Self-check answers and explanations
9. Appendix C. Exam day guide
10. Appendix D. Reference documents