Chapter template
Last reviewed: YYYY-MM-DD — see the freshness policy.
Every SA AI chapter should help the reader make and defend an architecture decision.
Learning objectives
Use three to five observable outcomes: select, design, compare, measure, defend. Avoid vague verbs such as understand.
Decision in one sentence
State the decision the chapter equips the reader to make.
Why it matters
Anchor the concept in the recurring Northstar case or another realistic scenario. Name the failure caused by getting it wrong.
Mental model
Explain the durable, vendor-neutral model. Use one compact Mermaid diagram when relationships or sequence matter.
Options and trade-offs
Compare at least two plausible options. Include quality, risk, latency, operability, and cost where relevant.
Architecture walkthrough
Trace inputs, trust boundaries, data flow, failure behavior, observability, and human responsibility.
Practical artifact
Produce a reusable SA artifact: decision matrix, C4 diagram, threat model, evaluation plan, cost model, ADR, or review checklist.
Lab
Link a public starter and state the exact deliverable and acceptance criteria. A chapter without a code lab may use a decision lab.
Check yourself
Ask three to five questions that require reasoning, not recall. At least one should apply the concept to an unfamiliar case.
Further reading
Prefer primary documentation, standards, and papers. Record a last-reviewed date and avoid undated claims about fast-moving products.