The Prompting Path curriculum

Prompting Path is a 30-day AI course in four weeks, plus a nine-module advanced track. Each day is one short lesson and one real task, about 10 minutes.

Day 0 to Day 8 are free, and each has its own page below. The other 21 days and the advanced track are a one-time £49 unlock. Of those, Day 14 and Day 21 are free previews, with their own pages too.

By Liam D, Founder and Principal Advisor at Futureformed · Updated 15 September 2026

Week 1: Foundations

How to talk to AI so it actually helps. About 70 minutes in total.

1. Day 0Pick your tool & meet the toolkitPick one AI assistant for the month, send your first message, and find the AI already inside the apps you use.10 minFree
2. Day 1Your first great promptWrite a prompt with the four ingredients that turn a vague request into a useful answer: role, task, context and format.10 minFree
3. Day 2Take it for a spinPlay with AI for ten minutes, away from work, to learn where it shines and where it stumbles.10 minFree
4. Day 3Show, don't tellUse few-shot prompting: paste two or three real examples so the AI copies your style and structure.10 minFree
5. Day 4Give it room to thinkAsk the AI to reason step by step before it answers, for better decisions, plans and analysis.10 minFree
6. Day 5When AI gets it wrongLearn where AI makes things up, which facts to always check, and how web search and stronger models help.10 minFree
7. Day 6Week 1 reflectionLook back on Week 1: what surprised you, what worked, and what you would now prompt differently.10 minFree

Week 2: Context is king

The biggest lever for better answers, and how to push back. About 70 minutes in total.

1. Day 7Stuff the context windowPaste in the full thread or document, learn what tokens and context windows are, and watch the answers improve.10 minFree
2. Day 8Long documents & summariesStructure prompts for long documents, and ask them better questions than a plain summary.10 minFree
3. Day 9Don't just ask, push backStart the judgment loop: form a rough view first, then work the AI's output in five modes.10 minFull path
4. Day 10System prompts & personasWrite custom instructions that tell your AI who you are and how you want answers, in every chat.10 minFull path
5. Day 11Formats & the tone dialSet the format and tone of an answer precisely, and find the tone that sounds like you.10 minFull path
6. Day 12Build your first templateTurn a task you do every week into a reusable prompt template you can find in five seconds.10 minFull path
7. Day 13Week 2 reflectionAnalyse the gap between your view and the AI's output: what it added, what it got wrong, and what only looks right.10 minFull path

Week 3: The right tool

Models, the AI in your everyday apps, files, images, memory. About 70 minutes in total.

1. Day 14Which model whenLearn the fast, balanced and heavy model tiers, pick your default, and know when to go heavy.10 minFree preview
2. Day 15AI inside your everyday appsUse the AI built into Word, Excel, Outlook and Google's apps for formulas, drafts and summaries.10 minFull path
3. Day 16Projects, Artifacts & CanvasSet up your first Project to keep instructions, files and related chats together, and meet Artifacts and Canvas.10 minFull path
4. Day 17Files, images & voiceUpload files, screenshots and photos, and use voice, for the tasks that are hard to type.10 minFull path
5. Day 18Image generation, with careKnow when an AI image helps, when it can mislead, and what to check before you use one.10 minFull path
6. Day 19Memory & external brainsLearn why AI forgets, and start an external brain of recap notes that you feed back in when needed.10 minFull path
7. Day 20Week 3 reflectionTake stock of your AI setup: your tool, your default model, the app features you use, and what you skip.10 minFull path

Week 4: From chat to workflow

Build something small that saves you time every week. About 90 minutes in total.

1. Day 21The repeat-tasks auditList your recurring tasks and rate them, to find the best candidate for your first AI workflow.10 minFree preview
2. Day 22Pick and scope your buildChoose one small task and write a one-page spec: input, output, rules, and what great looks like.10 minFull path
3. Day 23Build v1Build your first working prompt from the spec, run it on a real, messy input, and get it 70% right.10 minFull path
4. Day 24Iterate to v2Test your workflow on three real inputs, find where it fails, and fix it with examples and checks.10 minFull path
5. Day 25Make it reusableSave your workflow as a Project, Custom GPT, Gem or skill, so it runs the same way every time.10 minFull path
6. Day 26Connect a tool (if you can)Connect your AI to one tool you use, such as your calendar or email, and have it do a concrete job.10 minFull path
7. Day 27Use AI responsiblySpot trust drift, data leakage, bias and skill loss, and write your own three-rule AI policy.10 minFull path
8. Day 28Show your workFinish the judgment loop: share an AI-assisted piece with a short trail of what the AI did and what you changed.10 minFull path
9. Day 29Cross the finish lineReflect on thirty days, draw your map of where to trust AI, and claim your certificate.10 minFull path

Advanced track

Agents, your own toolkit, automation, dashboards, coding with AI, staying in control, efficiency, and your external brain. About 145 minutes in total.

1. A1Automate the routineSet up a scheduled AI task, such as a weekday morning brief, that runs without you.15 minFull path
2. A2Worked example: the book-club websiteBuild and publish a one-page website with Claude, GitHub and Vercel, with no developer needed.20 minFull path
3. A3From artifacts to dashboardsCombine artifacts into a live dashboard that answers the three questions you ask most about your data.15 minFull path
4. A4Build something real with Claude CodeDescribe a small piece of software in plain English, build it with Claude Code, then test it and fix it.20 minFull path
5. A5Your external brain, in practiceBuild a knowledge base and a Project loaded with your key documents, so every chat starts with your context.15 minFull path
6. A6How agents actually workLearn how an AI agent works in a loop with tools, and design two simple agents on paper.15 minFull path
7. A7Building your own toolkitClimb the ladder from saved prompts to custom assistants, connectors and skills, and build the next rung.15 minFull path
8. A8Staying in control as AI does moreSort AI tasks into three levels of human oversight, and audit the tools and connectors you have installed.15 minFull path
9. A9Why your AI session runs out of steamLearn why long chats slow down and hit limits sooner, and the habits that keep your sessions sharp.15 minFull path

Start with Day 0

No coding and no paid AI subscription needed. Pick Claude, ChatGPT, Gemini or Microsoft Copilot, and give it ten minutes a day.

Start free → See pricing