My New Favorite Prompt

After TechCon 365 Seattle last week—and a not-so-short but beautiful 1200-mile drive from Lehi, Utah to Farmers Branch, Texas over the weekend—I had some time to think about all the Copilot training I’ve been doing lately, and thought I’d write a quick post on the topic.

Game Play PromptThere’s one idea I keep coming back to: you don’t get better at prompting by reading about prompts—you get better by writing them. Over and over.

That’s why I’ve fallen in love with what I call the “Game Play” prompt. It’s hands down the best way I’ve found to help people build real prompt engineering skills—not just as individuals, but as teams.

What is a “Game Play” Prompt?

It’s exactly what it sounds like: a prompt designed like a game. You give the AI a role, a structure, and some rules, and it turns into an interactive coach.

Here’s a simplified version of mine:

We’re going to play a game involving prompt engineering.
You will give me a single task that can be solved with a prompt (no actual coding required). I’ll try to write that prompt. You’ll evaluate how well it solves the task, give me a score out of 10, and tell me how I can improve.

After every round, you’ll ask me if I want to try again. Keep going until I say “stop.”

Ask me the first question.

The AI takes it from there. It gives you a scenario. You write a prompt. It provides the output to your prompt and tells you what worked, what didn’t, and how to get better. Then it dares you to beat your last score.

Bonus tip: ask the AI “What would a 10/10 prompt look like for this task?” and let it show you what “great” actually means.

Why This Pattern Works

Most people still approach AI tools like Google: type a question, hope for a useful answer, repeat. That might get you part of the way there—but real AI fluency requires something deeper.

You have to move from search thinking to prompt thinking.

Prompt thinking is about designing your input like a blueprint—clear, contextual, and intentional. It’s a new skillset, and it doesn’t come automatically. Like any skill, it takes reps. That’s why game-like exercises like this are so powerful: they turn learning into doing—and they make the process sticky and fun.

The Skills Gap Is Real

Here’s the kicker: companies are already feeling the heat when it comes to AI adoption. According to a 2024 LinkedIn Workplace Learning Report:

  • 84% of global workers say they’ll need to learn new skills to keep up with AI.
  • But only 38% of organizations have formal training in place for AI tools.
  • Meanwhile, prompt engineering is one of the fastest-growing skill sets across tech and non-tech industries alike.

This is the new productivity curve—and those who know how to work with AI will have a serious edge.

Like I keep saying in trainings, and reminded my audience in my AI-Assisted Writing session at TechCon:

AI won’t take your job. But someone who knows how to use AI will.

Try the Game. Build the Muscle.

If you’re trying to get your team more comfortable with AI, start here. Let them play. Let them fail fast and iterate. Make prompt literacy something they build, not just study.

And if you’re ready to level up your skills…

👉 I’m launching a new online course all about Copilot and prompt engineering on August 1st.
We’ll go deep into practical strategies, hands-on exercises, and yes—plenty of game play. If you want to get early details before the landing page goes live, just shoot me a message or comment. I’ll make sure you’re on the list.

Let’s build the skills that future-proof your work.

Christian Buckley

Christian is a Microsoft Regional Director and M365 MVP (focused on SharePoint, Teams, and Copilot), and an award-winning product marketer and technology evangelist, based in Dallas, Texas. He is a startup advisor and investor, and an independent consultant providing fractional marketing and channel development services for Microsoft partners. He hosts the #CollabTalk Podcast, #ProjectFailureFiles series, Guardians of M365 Governance (#GoM365gov) series, and the Microsoft 365 Ask-Me-Anything (#M365AMA) series.