Teams buy “AI prompt training” and receive slide decks. The training host ships the opposite: click-and-do modules where learners run prompts, pass checks, and earn certificates—not passive slides.
In early July 2026, Prompt Anatomy shipped v1.4.2—the M7–9 Data Analysis path for Core buyers after the foundation modules. Stripe still sells Starter and Core only. M7–9 is an operator grant, not a checkout SKU. For hub checkout and the catalog split, see Shipping Prompt Anatomy. For the property map, start with The Prompt Anatomy Ecosystem Map.
How M7-9 access works
Who: Purchasers of the Core plan, Modules 1–6. See plans and pricing.
Offer: M7–9 as an operator grant. This is not a Stripe plan and not a public price change.
Request: Email info@promptanatomy.app with subject M7-9 grant, your purchase email, and LT or EN.
Starter-tier buyers, Modules 1–3, need Core first. The path assumes six-block fluency before data-analysis specialization.
Click-and-do beats slide-first
Decks teach block names. They rarely teach when to verify a number before it reaches a client. Analysts copy a polished output into email and dashboards, then invent metrics at scale. That failure mode is vocabulary without judgment.
Adults learn when they do something measurable in under three minutes. Duolingo and Udemy get the loop shape right. We borrowed the pattern, not the gamification: short sessions, copy-run-check, a certificate at a threshold, and a PDF handout for transfer. Streaks and leaderboards are not proof of maturity. The free workshop opener—map, assemble, diagnose—is Interactive Demos as Workshop Instruments.
| Pattern | Learner action | Learning aim | Where in training |
|---|---|---|---|
| Micro-loop | Choose, then get immediate feedback | Remember / Understand | M2 check, quiz gate |
| Copy-and-run | Fill blocks and run in a real AI tool | Apply | M3 business scenarios |
| Branching path | Pick an analytics focus and step through that diagram | Analyze | M7 core plus four thematic branches |
| Gated test | Reach 70% or higher to unlock certificate tier | Evaluate | M8 to Tier 3 certificate |
| Capstone | Produce a MASTER PROMPT and 8-step workflow | Create | M9 dominant path |
| Offline transfer | Download a PDF handout, not another slide | Transfer | M1, M5, M6, and M7-9 handouts |
Assembly stays in the drills. Execution stays in the model the learner already uses. That split is The Model Is Not the System.
What M7-9 covers
Modules 1-6 teach the six-block prompt system, the foundation every other path assumes. M7-9 is the data-analysis specialization: pipeline thinking, data preparation, visualization choices, anti-hallucination habits, and a capstone that forces an eight-step workflow instead of a one-shot summary.
M7 opens with a focus choice: visualization, ethics-plus, technique, or strategy. Learners move through a shared core plus thematic branches. v1.4.2 made those diagrams step-navigable and, on M9 completion, added a two-page Data Analysis handout in LT and EN.
M8 is the knowledge check that gates Tier 3 certification at 70% or higher. M9 is the capstone: a MASTER PROMPT workflow with extended scenarios for teams that want volume practice. Depth on grounding and retrieval stays on this site: Grounding AI Outputs and RAG in Production.
Soft launch is a readability and diagram-trust pass on the shipped path, not a promise that every interactive schema had full manual browser QA. What it does include is the focus branches, the M8 check, the M9 capstone, the 70% line, and diagrams that behave like instruments. Modules 10–12 are a later path for existing buyers, in 3A Before You Build an Agent.
A Tier 3 certificate proves practice inside the drills. The eval gate on a workflow you own still sits on your side. Use Evaluation Hooks for AI Workflows.
Part 1 — soft launch. Six org-role paths on Module 7, in v1.4.7, are in Role Paths Before Generic Analytics. Gated agent practice is 3A Before You Build an Agent.