You now have the framework, the patterns, and ten ready-made templates. This final lesson protects all of it. The difference between people who get steadily better at AI and people who start from zero every week comes down to one habit: saving what works.
Why Most People Don't Save Their Prompts
When a prompt works, most people copy the output and move on. The prompt itself disappears into chat history. Next week, the same task comes around again — and they face the same blank page, rebuilding from scratch what they already solved.
A prompt library changes that. Every good result becomes a template. Every iteration makes it a little better. After three months you have a competitive asset no one else has — your prompts, tuned to your work and your voice, sharpened by real use. The model is the same one everyone can access; your library is the part that's uniquely yours.
The setup takes about 30 minutes. The compounding starts immediately — the very next time you reach for a saved prompt instead of starting over.
Three Simple Folder Structures
You do not need special software. Pick whichever of these you'll actually open under deadline.
Notion (recommended)
Create a database with fields: Name, Category, Tool (ChatGPT/Claude), Last Used, Version, Notes. Filter by category when you need a prompt. Tag new versions instead of overwriting.
Google Docs
One doc per category. Heading 2 for each prompt name. Keep the prompt in a text block with [brackets] for variables. Add a short note on when to use it. Simple, searchable, shareable.
Plain Text File
If you want zero friction: one .txt file, prompts separated by --- dividers. Works everywhere, syncs via Dropbox or iCloud, never loses formatting.
What to Save vs What to Discard
A library is only useful if it stays clean. Save a prompt when it clears this bar — and let the rest go.
| Save This | Discard This |
|---|---|
| Produced professional output first time | Needed more than 3 fixes |
| You'll do this task again | One-off task with no repeatable pattern |
| Saves more than 10 minutes | Faster to write fresh |
| Consistent across different contexts | Too specific to one client |
How to Version Prompts
Prompts degrade. AI models update, your style evolves, client needs shift. A prompt that worked six months ago might produce worse results today — not because you changed it, but because everything around it did.
Simple versioning fixes this: add v1, v2 to the name. Keep the old versions — sometimes v1 was actually better, and you'll want to roll back. Add a date and a one-line note on what changed and why.
Review your library every 90 days. Retire prompts you haven't used. Promote the ones you reach for constantly into a "Favourites" section so your best tools are always one click away.
Build Your First Library in 30 Minutes
Don't plan it forever — build it today. Here's the whole thing in three timed steps.
Set up your structure (10 min)
Choose Notion, Docs, or plain text. Create the folders or database. Don't overthink the categories — start with: Writing, Research, Client Comms, SEO.
Save your first 5 prompts (15 min)
Go back through your recent AI conversations. Find 5 prompts that produced good results. Copy them in. Add the context note.
Add the Lesson 4 templates (5 min)
Paste the 10 templates from Lesson 4 into your library with the category tags. You now have 15 prompts to start from.
You've Completed the Course
You now have the framework (RTFC), the techniques (few-shot, chain-of-thought), 10 ready-to-use templates, and a system to build on all of it. Next step: apply one prompt from your new library today.
Your prompt library is not a collection of tricks. It's a record of what works for your work, in your voice, for your clients. Start it today. It compounds from day one.
Continue Learning
Next course: AI Tools Fundamentals — learn which tools to use your new prompts with.
Then: AI Automation with n8n — build workflows that run your prompts automatically.