Typing in a simple question typically won’t get you the results you want from an AI model.
Just as humans often go through multiple drafts before finalizing their work, your AI model may need some guidance to produce the best output, according to Jordan Wilson, the founder and host of Everyday AI, a top-20 tech podcast and AI strategy company.
Wilson, who is also a professor of AI at DePaul University in Chicago, says that he teaches his students to use a process called “prime, prompt, polish” to improve their AI results.
The purpose of this framework is to help break people out of the mindset that they “get one input and one output out of a large language model,” he says in Smarter by CNBC Make It’s course, Beyond the Basics: How to Use AI to Supercharge Your Work.
Instead, Wilson recommends engaging in a back-and-forth dialogue with the model.
