AI Prompt Tip of the Day Collection

Creating a library of prompts that get the best responses

As you know if you’ve spent any time looking at Tech & Nonsense, I’m an AI Optimist, an “always invite AI to the Table” sort of person. I use four GenAI apps on a daily basis and a couple others here and there. I’ve tried out a few different ways of keeping track of prompts to use with these apps that get the best responses. Multiple notes in Notion and Obsidian are on the topic of ‘Prompt Engineering’ and many are long lists of recommendations on the best ways to prompt these tools. I also have a Notion database (that I haven’t added much to lately) that tracks prompt responses by type and by Great/Good/Weak responses.

The one that has been both my most used and most effective over recent months is Notion note called ‘Prompt Tip of the Day Collection’. It’s not super long yet, and I’m going to try to keep it that way by doing regular housekeeping on it. I’ll look to remove any of the prompts that are getting diminishing returns and add new ones as and when they bring good results.

Before I share a few examples, I need to give a bit hat tip to The Neuron and to Allie K. Miller, my primary sources for finding these. And to Ali Abdaal for this great foundational AI Prompts Cheat Sheet infographic:

Here are a few that are current favorites and most often used in my fledgling collection:

Triple-Vision Translator

“Explain [complex concept] three times: (a) to a 12-year-old (b) to a college student (c) to a domain expert who wants edge-case caveats”

The “Triple-Vision Translator” hack unlocks clarity at every level. Ask ChatGPT or Claude to explain any complex concept three different ways—for a six grader (or kindergartener, if you want a really simple version), a college student, and a domain expert.

Just copy-paste: “Explain [complex concept] three times: (a) to a 12-year-old (b) to a college student (c) to a domain expert who wants edge-case caveats”

The two extremes have been working best for me - explain it to a 12 year old and to a domain expert.

Spec Sheet Squeeze

The “Spec Sheet Squeeze” technique transforms generic AI responses into tailored gold. Instead of asking for “a marketing plan,” ask for “a marketing plan with these exact specifications” followed by a bulleted list of your requirements. We tested this on Claude and ChatGPT this week—the difference was night and day.

Just structure your prompt like this: “Create [deliverable] that includes: • [specific element 1] • [specific element 2] • [formatting requirements]”

The more detailed your spec sheet, the less editing you'll need to do afterward. When we added “must fit on a single PowerPoint slide” to our executive summary request, the AI actually delivered something presentation-ready on the first try!

I’ve used this one at home and at work, and it lives up to that description.

This one is excellent when brainstorming with these AI tools:

Let AI ask you questions

Instead of just asking AI questions, allow AI to ask you questions. AI can teach you how to use itself, unlike traditional tools like Excel or Powerpoint. (his sample prompt: Hey, you are an AI expert. I would love your help and a consultation with you to help me figure out where I can best leverage AI in my life. As an AI expert, would you please as me questions. one question at a time until you have enough context about my workflows, responsibilities, KPIs and objectives that you could make two obvious recommendations for how AI could leverage AI in my work.)

And these two are very helpful for challenging my thoughts or ideas:

Thoughtful Contrarian

“Here’s what I believe: [insert opinion]. What’s the strongest argument against it (slash help me “steel man” the other side)? How would a smart, skeptical person respond?”

What might I be missing?

Example: "Our project keeps getting delayed despite our best efforts. What might we be missing?"

Or better yet, have it brainstorm “at least 15 things I might be missing, ranked from most obvious to most unexpected, in a bullet point list. If there are less than 15 potential missing factors, include as many elements approaching 15 as there are.”

I’d love to hear which GenAI tools you’re using, and which types of prompts get you the best responses.