Editor’s note: This article by artificial intelligence expert Nicole Donnelly is the fourth in a series appearing in SAF News Now about how to use AI for your business.

Ever feel like the artificial intelligence platforms you use just don’t understand what you are asking them to do? There is a way to solve that problem: Think like an engineer.

Prompt engineering is a term for asking AI to do what you want it to do, and it has revolutionized how we use AI. It’s the art and science of “talking” to AI and asking AI the right questions in the right way to get the outputs you want.

Here are nine tips to help you master the art of prompt engineering.

1. Understand the concept. Prompt engineering is about designing input queries or prompts to elicit the desired output from a language model like AI. Think of it as a new version of library science.
2. Be precise. The more precise your prompts are, the better the AI can assist you. Clarity is crucial, so be explicit with your requests. Provide specifics, context, simple and concise language and give examples and ideas in your prompts. For example, rather than ask, “List 10 names for a flower arrangement,” it is better to ask, “List 10 names for a red, white and green flower arrangement made up of roses, gerbera daisies and carnations that is sold by a florist for Christmas.”
3. Understand limitations. Every AI platform has its limitations. It may not have the most current real-time data or grasp the latest cultural nuances, which could affect its responses. It may also not be able to retain all of the information that you put in earlier. And, always double check the outputs — even AI can mess up simple math.
4. Craft niche-specific prompts. When asking for marketing strategies, be specific about your industry or niche. For instance, ask about digital marketing strategies for a flower shop or farm rather than a generic question about improving your marketing strategy. For example, rather than asking for a digital marketing strategy, ask it for a digital marketing strategy that will increase pre-orders of flower arrangements in the weeks leading up to a floral holiday.
5. Experiment with different AI platforms. Different versions of AI tools will yield different results. Don’t shy away from trying and comparing them. We often test the same prompts on Bard, ChatGPT, and Claude2 weekly to see which platform does best on various types of requests.
6. Upgrade if necessary. If you’re using the free version of an AI tool, consider upgrading to a paid account for better results. ChatGPT 4 is 10 times smarter than the earlier version.
7. Use tools for efficiency. Artificial intelligence-powered response manager (referred to as AIPRM), enhances the user experience by providing a community-driven prompt library. Download the AIPRM tool of your choice online, and use the pre-written prompts (and write your own prompts for things that you do often, such as writing job listings).
8. Reiterate and expand. If you want more information on a certain section of your AI-generated document, don’t hesitate to ask. The AI can expand on any part you specify. For example, if you really liked an idea for a social media campaign, ask it to provide a timeline for the campaign and provide ideas for images, captions and hashtags. This iterative prompting tends to be the most effective. For best results, make your requests in small bits, then ask for a revision, then another, until you get the version that you want.
9. Keep an open mind. The possibilities with prompt engineering are endless. Experiment with different prompts and see how the AI responds.

Remember, the goal of prompt engineering is to turn a literal-minded AI into a strategic partner attuned to your specific needs. By iterating your prompts, you can leverage the power of AI across your organization for copywriting, data analysis, spreadsheet creation, image creation and editing, and more.

Nicole Donnelly is the founder of AI Smart Marketing, which specializes in various services including AI smart marketing strategy, implementation, website design and build, social media, and AI training.