Forum Discussion
Creating a Prompt Library
We have about 20 or so folks across the organization using Articulate 360. I want to start a prompt library that we can all contribute to for code block prompts, AI images, etc.
For those of you who have done something similar, what has worked best for you?
4 Replies
The problem I've found with a prompt library is they tend not to be used when they're just a wall of pasted prompts. Each entry needs: what it's for, the prompt itself with the variable bits marked, which tool it goes into (Rise code block, Storyline JavaScript trigger, AI Assistant image, AI Assistant text), an example of the output, and a note on what to change. It can get messy really quickly.
We've got around this by using the prompt library tool at the link below. It's 100% free to use. I'd advise reading the built in guide first. If you want the file to host your own version of the tool somewhere else, I'm happy to share. It will work in SharePoint or any other shared repository. It's a standalone HTML file. Your prompts stay in your browser; the only network request is for the fonts.
Thanks. I will take a look at this.
Cheryl — I've helped run one of these on an innovation team, and DShaw is right: the structure is the whole game. A few things I'd add from watching one survive past the first month:
1. Organize by task, not by tool. People think "I need a better quiz question" or "I need a hero image," not "I need a Rise code block prompt." If the library matches how people think about their work, they actually browse it.
2. Seed it yourself first. Launch with 10–15 prompts that already produce good output, example included. An empty library that asks 20 people to contribute on day one stays empty — nobody wants to be the first to write in it.
3. Date-stamp everything. Prompts go stale as models change — what worked in January produces different output by June. A quarterly pass to retire duds keeps the library trustworthy. A library nobody trusts is a library nobody opens.
A shared doc or wiki is honestly enough for the first version. DShaw's tool looks solid for later, but tools don't fix an empty or stale library — the habits do.
