“A classic argument, something I believe deeply to my core, is that new technology only ever increases the amount of jobs available.”
— Mike Taylor
Ever wondered what a “prompt engineer” does? I recently sat down with Mike Taylor—who quite literally wrote the book on prompt engineering—to talk about how we can all leverage AI in a more effective, human-friendly way. If you’ve dabbled in ChatGPT or other AI tools and found them “just okay,” read on. A few tweaks in how you talk to your AI can make all the difference.
Meet Mike: From marketing pro to prompt engineer
Mike’s background might sound familiar: he led a growing marketing agency, managed a 50-person team, and embraced remote workflows. Then, in 2020, he discovered GPT-3 (a precursor to ChatGPT), and—like many of us—fell headfirst into the rabbit hole of generative AI. Fast-forward to today: Mike’s authored Prompt Engineering for Generative AI, helps businesses streamline AI workflows, and firmly believes we’re only scratching the surface of AI’s possibilities.
Yet, it’s not just about fancy technology or new job titles. For Mike, getting good at prompt engineering is a lot like hiring a highly enthusiastic intern who needs the right context to deliver quality work.
“If you only spent five minutes handing an intern a task, you can’t expect miracles. It’s the same with AI—you have to teach it the context you already have in your head.”
Why prompts are everything
So what exactly is prompt engineering? In short, it’s the art (and science) of crafting inputs that get you the best outputs from AI models like ChatGPT, Claude, or Bard. If you’ve ever typed a single line into ChatGPT and felt underwhelmed, you’re not alone. Most default prompts yield generic text—“streamline,” “delve,” and “synergy” ring any bells?
Mike’s advice? Give your AI the same info you’d give a new hire. For instance, if you’re writing a blog post:
- Show Samples of Your Voice and Style. Provide a short piece you’ve written before. AI models can mimic that tone in future outputs.
- Break Down Tasks. Rather than asking for an entire blog post in one shot, start with an outline, refine it, then move on to the opening paragraph, and so on.
- Load Up on Context. The more relevant details you include (target audience, specific facts, or interview quotes), the better your AI can customize content.
“Every system in production splits the task into smaller sub-tasks,”
Mike explains. By handling each step one at a time, you end up with a stronger, more on-brand result.
Yes, it’s worth the effort
If you’re thinking, Wait, that sounds like a lot of work!—it can be, initially. But consider the payoff:
- Scalability: Once you’ve nailed the prompts for, say, writing blog posts, you can generate 10…or 200…or 4,000 pieces of content, far faster than a single human could manage.
- Consistency: A well-engineered prompt ensures your AI nails the style, structure, and accuracy you need—every time.
- Creative Partner: Tools like Claude, ChatGPT, or Bard can serve as brainstorming allies, helping you try new directions you might never think of on your own.
Mike points out that if you rarely do a certain task, you don’t have to go full throttle into advanced prompting—just keep a few best practices in mind. But if you repeat tasks daily or weekly (like coding, writing, or data analysis), refining your AI workflow is a game-changer.
Takeaways (without the AI “slop”)
- Treat AI Like a Collaborator
Approach ChatGPT or Claude as if they’re team members. Provide them with background, tone guidelines, and relevant resources. - Iterate in Steps
Outline, then bullet points, then draft paragraphs—this eliminates the one-and-done (often mediocre) approach. - Decide Where to Go Deep
You can’t master every AI topic. Choose one area—like marketing content, coding helpers, or scheduling—and refine prompts there. - Keep Learning, But Selectively
AI news moves fast. Follow a few experts on social channels, find your niche (e.g., multi-agent systems), and rely on your network to surface major breakthroughs.
Wrapping up
Prompt engineering isn’t rocket science, but it does require some practice.
For day-to-day tasks, adding even a pinch of structure can boost your AI’s productivity tenfold. Whether you’re writing a blog post or trying to parse massive data sets, small changes to your prompts can yield huge benefits.
Ready to give it a shot? Experiment with your next AI prompt, add some context, share a writing sample, or break down your request into steps. If you want to dive deeper, check out Mike’s book Prompt Engineering for Generative AI, or follow him on X (formerly Twitter) at @hammer_mt.
And, of course, if you’re balancing mountains of tasks across email, Slack, and meetings, Hoop can help you capture and prioritize them automatically—so you have more space to focus on what really matters. Because at the end of the day, it’s about making all our tools work for us, not the other way around.
