Decide what deserves to be an agent before you build one.

Module 00 of 16 · track Think · Estimated time: 4–5 hours · Prerequisites: None. Start here even if you have never called an LLM API.

The most expensive mistake in agentic AI is building an agent for a job a for loop could do. This module gives you the vocabulary and the decision tools to tell the difference, and the artefacts to defend that decision to a stakeholder.

Open the module on GitHub ↗: the slides, notebooks, exercises and solutions are there. This page is the reading order, with the tutorial and the workbench alongside.

The lesson that frames this module is P0 Frame: How to Decide If an AI Agent Is Worth Building (8 min): the phase this module's decision tools belong to. Then try the AI-fit assessor in the workbench: the same decision, with your numbers.

• Classify any candidate use case as script, workflow, or agent — and justify it
• Name the six recurring failure patterns of agentic systems and spot them in a design
• Write an invariance test that proves a behaviour is stable across paraphrases
• Produce a one-page agent PRD that a reviewer can actually challenge
• Estimate token cost per interaction before a single line of code is written

Portable GenAI concepts	AWS-specific surface
Agent vs workflow vs script	None yet — this module is deliberately cloud-free
Autonomy dials
Failure taxonomy
Invariance testing
Token economics

Portable concepts transfer to any stack. The AWS column is where this module touches the cloud — see docs/concepts/portability-matrix.md.

#		Step	What it is
1	📖	slides/Day1_Half1.pptx	Framing: what an agent is, and the four-quadrant classifier
2	📊	activities/H1-01_Four-Quadrant_Classifier.xlsx	Classify your own use cases in the workbook
3	📊	activities/H1-02_Is-It-An-Agent_Vote.xlsx	Team vote — argue the borderline cases
4	📊	activities/H1-04_Six-Failure-Pattern_Diagnostic.xlsx	Diagnose the six failure patterns
5	✏️	exercises/Exercise_Mid-Session.pdf	Mid-session checkpoint
6	📖	slides/Day1_Half2.pptx	From idea to PRD: scoping, guardrails, readiness
7	📊	activities/H2-01_PRD-Builder.xlsx	Draft your agent PRD
8	📊	activities/H2-03_Token-Cost_Calculator.xlsx	Cost your design before you build it
9	✏️	exercises/Day1_Half2_Applied_Exercise.md	Applied exercise — full design pass
10	✏️	exercises/Architecture_Exercise.md	Architecture exercise

📖 read   💻 run   ✏️ practise   📊 workbook   🔖 reference

Worked solutions live in solutions/. Attempt every exercise closed-book first — the solutions are written to be read after you have a wrong answer to compare against.

• Calling something an agent because it uses an LLM. Autonomy is the test, not the model.
• Writing outcomes as 'understand X'. If you cannot test it, it is not an outcome.
• Deferring cost estimation until after the build. By then the architecture is fixed.

Every error anyone has hit in this course is in the Error Index, by the exact string. Model access and region problems are in Model and Region Notes. If it is a question rather than an error, check the answered ones first, then ask in Q&A.

Frameworks

• Autonomy Ladder — Build the lowest rung that passes your acceptance test
• Scope Fence — Four posts that stop "just add it to the prompt" creep
• Evidence Ladder — Six rungs of proof, and the claim each licenses
• Value Trace — Five links from model metric to money

Recipes and procedures

• Agent design review — Running the review this module prepares you for

---

All modules · Module 01 · LLM Intuition and the AWS Bridge →

_{✏️ This page is generated from modules/00-agentic-foundations/README.md. To change it, edit the module README: an edit made here is replaced at the next sync.}