How to Use the RISEN Framework for Better AI Prompting

If left as is, AI will tell you what you want, albeit in the very literal sense. It’s easy to misdirect it or send it down rabbit holes that eat up significant time and money for your business. If you want to make full use of AI, reduce hallucinations, and be more productive with these incredible new tools, you need to master the art of prompt engineering.

Here are some best practices to consider for your AI prompting.

The RISEN Framework

AI needs a clear set of instructions, especially if you want to get anything of substance out of it. After all, the output can only be as good as the input. Use this simple rubric to achieve maximum clarity and specificity:

The RISEN Framework Explained

As you might expect, RISEN is an acronym that outlines the following steps:

“R” is for Role - Define the perspective that the AI is adopting with this prompt. If you want it to be a business consultant, then tell it that…but don’t stop there; get as specific as you can. So, for example, you might include a specific niche or expertise in your prompt. Ensure that you identify who the AI is emulating and an appropriate response might follow.

“I” is for Instructions - Next, define what you want your AI to do. This should be pretty self-explanatory, but the more detailed you can make these instructions, the better a response you’ll get. For more context, think about giving a machine a set of instructions (which is basically what you’re doing here). With more details, it’ll be less likely to stray from its course.

“S” is for Specifics - Remember, you have a specific idea for what you want from the AI, but the AI doesn’t know that unless you tell it what you want. For example, if you’re in an industry where compliance is mandatory, you want to mention that. The last thing you want is to be in violation of something like HIPAA or NIST. With enough detail and guidelines in place, it’s only a matter of time before you get exactly what you’re looking for.

“E” is for Example - Here’s where most people drop the ball. Instead of simply telling the AI what you want, show it through an example. Use an existing report or template so it can make content that matches what you need.

“N” is for Negative Constraints - Yes, the AI needs to know what to do, but it also needs to know what not to do. These negative constraints can help ensure the AI delivers exactly what you want, and nothing more.

The Key is Context

Remember that the AI only knows what you disclose to it, so you’ll have to provide information such as the audience, your goals, its target voice, and plenty of grace. It might take multiple attempts before it gets to a place where you’ll find its output acceptable.

One disclaimer: You should never, ever, provide public AI tools with sensitive or proprietary data. These systems work by sharing data, so if you provide it with anything proprietary, it can use that to build a response for someone else… even your competitors.

Adjusting the Tone

Some AI tools also offer ways to change the language used by AI. For example, you can set it to provide creative responses rather than critical or factual responses, allowing you to effectively change the “temperature” of its text.

Want to Get More Out of Your Business?

We can help your business use AI, along with countless other IT solutions, to improve operations. Call us at (276) 601-3208 to learn more.