Agent Workflow Foundations

Lessons in This Module

1. Lesson 1

   Objectives, Decisions, Actions, and Approvals

   An agent workflow separates the outcome being pursued from the judgements made, the actions available, and the approvals required before consequential work proceeds. Learn the founder decision, practical evidence, and warning signs that matter.

   Open Lesson

2. Lesson 2

   Context, Tools, Constraints, and Expected Outputs

   A bounded agent responsibility combines relevant context and approved tools with explicit constraints and a reviewable definition of a satisfactory output. Learn the founder decision, practical evidence, and warning signs that matter.

   Open Lesson

3. Lesson 3

   Deterministic Automation or Agent Reasoning

   Deterministic automation follows stable rules, while agent reasoning interprets variable information and chooses among bounded options. Learn the founder decision, practical evidence, and warning signs that matter.

   Open Lesson

4. Lesson 4

   Human Review Proportional to Risk

   Proportional review increases human oversight as consequence, irreversibility, uncertainty, or sensitivity increases. Learn the founder decision, practical evidence, and warning signs that matter.

   Open Lesson

5. Lesson 5

   Testing Exceptions and Failure Cases

   Agent evaluation uses representative tasks, missing information, conflicting evidence, unavailable tools, unsafe requests, and recovery cases rather than one prepared demonstration. Learn the founder decision, practical evidence, and warning signs that matter.

   Open Lesson

6. Lesson 6

   Maintaining Agent Workflows Over Time

   Agent workflow maintenance keeps instructions, tools, permissions, evaluations, owners, and operating assumptions current as the surrounding system changes. Learn the founder decision, practical evidence, and warning signs that matter.

   Open Lesson