Make the Most of Prompt Engineering: A Guide for Writers

Most of us are now at least familiar with the concept of AI tools, although many of us are still wrapping our brains around how to put them to good use. To start, we can use prompt engineering, which is the tool guiding generative AI models (like ChatGPT, Bard, Bing, Claude, and an endless stream of others) to provide us with text, imagery, or other desired outcomes from what we feed it.

There are dozens upon dozens of prompt engineering tools capable of a vast array of tasks, especially for those of us in the content-producing biz who want specifically tailored text responses to solve our copywriting problems. These “problems” could mean crafting targeted ad campaigns, writing in-depth reports, redoing resumes and cover letters, and generating lists of all types. Anything we may need to write, prompt engineering can help us get there just by loading very specific prompts in a language model (like ChatGPT) to produce “high-quality, targeted content with minimal effort.” 

Again, the main premise (and promise) is the more well-crafted our inputs are, the more specific and high quality the outputs. This dream sure seems possible, especially given how fast some of these tools have already evolved in a short period of time, getting smarter with each iteration. But there continue to be major concerns, both potential and real. Should we embrace these wonder tools? Will they free up humans to concentrate on other things while quickly churning out the copy we need? Or will such tech magic put us all out of business (especially the biz of writing)?

Marketers regularly tout the benefits of generative AI and prompt engineering, and many content creators boast of its capabilities. Along with this big sell, however, some of the sharpest thinkers around—Steven Pinker, cognitive psychologist / psycholinguist, and Jakob Nielsen, UX pro, to name only two—caution users on volatility, ethics, and false claims inherent in these seemingly wondrous tools. 

Is prompt engineering the game changer folks claim?

Along with other copywriters and content producers, I of course gave this nifty tool a whirl early on—and still engage it for various tasks: creating quick-hit lists and comparing outlines, drafting short and sweet paragraphs from dense descriptions, generating quippy headlines, and occasionally taunting it into crafting specifically formatted love poems, haikus, and horoscope profiles (all epic fails). While it’s been interesting to see how tools like ChatGPT and Bard respond to the same prompts differently, and continue to evolve, I went into this being dubious about quality, creativity, and ethics—to name just a few colossal concerns—and remain staunchly in this camp. 

Prompt engineering allows writers to take advantage of AI’s strengths—efficiency and productivity, mostly speed. This is helpful when facing writer’s block. Yet what about minimizing its inherent weaknesses? These content-generating tools continue to present unfactual information as fact, and confidently so, and consistently offer a pervasively bland, flat, boring approach and tone.

Even ChatGPT is on to its faults, spitting out: “… developing a brand voice, creating a content strategy that aligns with business goals, and crafting engaging stories are all areas where human input is crucial. Content strategists and copywriters also bring a deep understanding of their target audience, which cannot be replicated by AI…Moreover, AI-generated content may lack the emotional resonance and nuance that humans are able to convey. Humans can tap into their own experiences and emotions to create content that connects with readers on a deeper level. AI, on the other hand, is limited by the data and rules that it has been trained on.”

Does prompt engineering generate better writers and thinkers? 

Many thought leaders believe that AI won’t ever compete with the nuanced and creative human mind, but that very mind (ours! mine!) can steer prompt engineering into producing high-quality, high-interest, highly credible content just by asking it the right questions. The more thought we put into our prompts, the more likely we’ll get copy optimized for specific calls to action, conversions, and further engagement. Asking questions, and then refining those queries to get better quality responses, is never a bad exercise. 

I’ve been steadily refining my approach to my ChatGPT buddy, formulating better and tighter prompts for each unique task at hand, evaluating the output, and using that to move my own human writing forward. The more detailed and specific my request, the better quality IS its response. I enjoy noting progression with each refinement, both human and robot. 

Here’s a best practices approach that’s been working for me (and many other writers using prompt engineering tools):  

✓ Know your objective: Set a goal for your intended outcome, and provide those goals as part of your query.

✓ Be specific: Guide AI toward analysis and detailed scenarios.

✓ Be concise and direct: Use simple, clear, descriptive language. 

✓ Designate limits: Provide format and time-frame constraints.

✓ KIS (Keep It Simple): Avoid jargon, use quotes, and set parameters to guide best outputs.

✓ Rinse and repeat: After reviewing an output, refine your query and feed the machine again.

Some of my college professor pals in U.S. and European universities are challenging their students to use AI and prompt engineering ethically and with best practices, since there’s no avoiding this tool in academic settings. Comparative critical thinking and writing exercises reveal that students KNOW they can produce better copy than what AI is spitting out (yay!), that the machine’s churned-out writing is uncreative and uninteresting. So, students try it out, evaluate the results, refine their questions to produce stronger writing. Uptick: Prompt engineering as a critical thinking and writing tool works!

Moving forward with a faulty tool

For me, it’s been both interesting and affirming that for the specific task of generating creative writing and fresh ideas, AI ain’t all that. But using this tool HAS made me a better question asker, direction honer, goal focuser, and contributes to smarter, more targeted answers (ChatGPT’s), and tighter, more creative writing (mine, hopefully). 

Is the robot getting more attuned to our relationship? I get the feeling it senses my just-under-the-surface anxiety: “While AI can certainly assist with certain aspects of content strategy and copywriting, it is unlikely to completely replace human expertise in these areas. Instead, AI should be viewed as a tool that can help augment human creativity and efficiency, rather than replace it entirely.”

WHEW. 

[Author’s note: Prompt engineering tools were neither used (except for the 2 ChatGPT quotes) nor harmed during the crafting of this blog post.]