Ever ask ChatGPT for something and get a super generic, useless answer? Yeah, I’ve been there. It’s so frustrating!
But I just stumbled upon an awesome post from an industry pro who points out that the problem often isn’t the AI, it’s our prompts.
The person who shared it laid out some incredible prompting frameworks to give the AI a proper brief. I think this is a total game-changer for getting better, more specific results.
Here are a few of the frameworks the author shared:
- ✅ TREF: What’s the Task, what are the Requirements, what do you Expect, and in what Format?
- ✅ SCET: Describe the Situation, the Complication, what you Expect, and the final Task.
- ✅ PECRA: Explain the Purpose, your Expectation, give some Context, make a Request, and suggest the Action.
- ✅ GRADE: Tell the Goal, make a Request, specify the Action, add Details, and give Examples.
- ✅ ROSES: Mention the Role, the Objective, give a Scenario, share the Expected Solution, and list the Steps.
By structuring your request like this, you’re giving the AI all the context it needs to deliver exactly what you want. So simple, yet so effective!
The original post from this talented creator has even more frameworks and a cool infographic. Definitely go read the full post to level up your prompting skills!