Prompt Anatomy is a methodology for predictable AI operations, not a collection of clever prompts. This page routes you into the blog by job-to-be-done. If you need to know which property to use first—blog, training app, or marketing site—read the Ecosystem Map before picking a path below.
The stack
The stack is the reading order for implementation—not a shopping list of models and copilots. Outcome and workflow come first; the model step sits inside context, eval, and governance. Skipping layers produces demos that cannot survive audit or scale.
Business outcome
→ Workflow (steps, owners, handoffs)
→ Context architecture (what the model may see)
→ Model step (generation where it adds leverage)
→ Evaluation (pass/fail before scale)
→ Governance (who may change what)
Deep dive: The Model Is Not the System.
Where to go next
Use the table as a routing index by job-to-be-done. Pick one row, finish that playbook, then return here—foundations is a hub, not a reading assignment for every stakeholder on day one.
| If you are… | Start with |
|---|---|
| Diagnosing chaos | 10 Signs Your Company Is Vibe Prompting |
| Designing agents | Agent workflow guide — path visual: From Prompt to Agent |
| Designing context | Context architecture |
| Setting ownership | Governance roles |
| Planning a pilot | AI Workflow Canvas |
| Freezing tool spend | AI Procurement Freeze |
| Proving ROI | Measuring AI Workflow ROI |
| Running eval gates | Evaluation Hooks + Eval Checklist |
| Picking automation | Workflow automation selection |
First 30 days
Follow this sequence once—not every playbook on day one.
- Diagnose with 10 Signs and the stack scorecard.
- Design one workflow on the AI Workflow Canvas.
- Gate changes with Evaluation Hooks before pilot scale.
- Standardize on training at promptanatomy.app when the team repeats the same steps.
For outcome framing before you buy anything, read From Prompts to Business Outcomes.
By persona
Operations or enablement (diagnose first). You need a shared vocabulary for what is broken before anyone buys another copilot. Start with 10 Signs and the 15-minute stack scorecard; use Glossary terms in steering meetings so “eval” and “workflow ID” mean the same thing to IT and the business. When tool trials are the blocker, read AI Procurement Freeze before the next vendor demo.
Architecture or engineering (design the system). Map one workflow on the canvas, wire Evaluation Hooks before connectors, then read Grounding AI Outputs and RAG in Production. Agent paths: How to Design an AI Agent Workflow and MCP for enterprise. For maturity framing, see the AI Implementation Maturity Ladder.
Governance or risk (control change). Pair Governance roles with the Governance RACI Worksheet, Audit trails, and Risk review cadence. Prompt and context changes ship through the Prompt Registry Playbook—not shared docs.
Templates and proof
Templates and case studies turn methodology into artifacts auditors and sponsors can inspect. Copy them into your wiki with workflow IDs filled in—empty templates in a shared drive do not count as governance.
- Templates — Eval checklist, RACI worksheet, change log template
- Case studies — Northline Part 1 (diagnostic arc); finance workflow for controlled draft and review
- Browse all — Templates and Case Studies
Training and hands-on drills: promptanatomy.app. Term definitions: Glossary.
Framework overview article retired—this page is the single framework index.
