Course 02 · Beginner to intermediate · 6–7 hours · Learn at your own pace
Writing powerful, reusable prompts
Design, test and document reliable prompts, then bring them together in a ready-to-use library.
Complete course, currently in quality review before public opening.
- Estimated duration
- 6–7 hours · Learn at your own pace
Estimate calculated from the content (reading and exercises), not measured with learners: modules ≈ 3 h 45 min, assessment ≈ 5 min, final project ≈ 2 h 40 min.
- Level
- Beginner to intermediate
- Prerequisites
- Having already used a conversational AI assistant (ChatGPT, Claude, Gemini, Copilot…).
- No programming skills: the JSON basics you need are explained in module 4.
- Required tools
- A recent browser: every exercise runs in the course pages, with no AI and nothing to install.
- Optional, for the final project: an AI assistant of your choice to run your prompts (a free plan is enough).
- External costs
- No mandatory external cost. Exercises are checked in your browser; no API key is ever requested.
Learning objectives
- Diagnose why an instruction produces a vague or invented answer.
- Structure a prompt in six components: objective, context, inputs, constraints, output format, verification.
- Turn a one-off request into a reusable template with documented variables.
- Require a usable format (table, checklist, JSON, report) and fix a non-compliant output.
- Write test cases and compare two versions of a prompt objectively.
- Deliver a library of five documented, tested prompts for a real workflow.
Syllabus
-
Module 2 — Structure a professional prompt
The six components of a reliable prompt, and a builder to assemble them.
-
Module 3 — Variables and reusable templates
Turn a one-off request into a template the whole team can reuse.
-
Module 4 — Controlling output formats
Tables, checklists, JSON and reports: require a usable form and fix a non-compliant output.
-
Module 5 — Testing and evaluating a prompt
Test cases, observable criteria and an objective comparison between two versions.
-
Module 6 — Build a professional prompt library
Document, chain and maintain prompts — then complete the final project.
Final project: the customer service prompt library
A library of 5 prompts for a fictitious company's customer service: templates, variables, examples, expected outputs, test cases, checklist and documented improvements — assessed with a published rubric and exportable as Markdown and JSON.
Downloadable resources
- Prompt sheet (blank template)prompt-sheet-template.md
The documentation structure of a library prompt, to copy for each of your prompts.
- Test plan (spreadsheet)test-plan-template.csv
A CSV sheet to record your test cases and their results, version by version.
- Final project rubricfinal-project-rubric.md
The five criteria, their levels and the pass rule — identical to the online assessment.
- Sample customer messages (fictitious data)sample-customer-messages.json
Ten fictitious Atelier Lumen messages to test your prompts: common cases, edge cases and invalid inputs.
- Reference solution: “Message triage” sheetreference-solution-triage.md
A complete sheet at the level expected for the final project (one prompt only: the other four are yours).
- Blank project to importlibrary-project-blank.json
The final project's JSON structure, to fill in offline and then import into the project editor.
Content verified on 28 October 2026
Practical training provided by Cloud Tech Innovation. It awards no diploma or recognised certification; the summary only records the exercises passed and the assessed project.