|
Most people think the magic of AI is in the answers. But the real magic? It’s in the questions — and how you write them. Because let’s be honest. Asking AI for something then getting a sloppy answer really kills the mood. Makes a person think it’s not worth the hype. But who has time to research and learn “effective prompting” from the army of AI bros on social media? Not me. So here’s the shortcut almost nobody talks about: You can use AI to help you talk to AI. I know, it sounds a bit like standing between two mirrors — but that’s the point. When you ask ChatGPT to “make this prompt better,” it doesn’t just rewrite your words. It exposes what the AI actually needs from you — details, tone, format, constraints. It’s like peeking behind the curtain of how the system thinks. Say you want to generate a product image in Veo 3, or create a character concept in Midjourney. Instead of typing a vague line like “make a cool cyberpunk wizard”, you can drop into ChatGPT first and ask: “Write me a detailed prompt for Veo 3 that will create a cinematic video of a cyberpunk wizard walking through neon rain. Include camera angles, lighting, and pacing.” If you're not even sure what to ask then ask ChatGPT what you should be asking or if there's anything others have used to make their prompts better. Not everyone is even going to think of angles and lighting for image prompt without experience, which you can gain through asking ChatGPT. Now you’re not just guessing — you’re leveraging one AI’s language skills to amplify another’s creative output. It works in the other direction too. You can ask Claude to refine your ChatGPT system prompt. Or use ChatGPT to help you build better instructions for automation tools like n8n, Make, or even a future agent builder. That’s meta-prompting — using one model to sharpen your thinking before you even start. Check out a quick example of me creating one here for YouTube And the more you do it, the more you realise: It’s not about replacing creativity with AI. It’s about replacing hesitation with structure. The question isn’t “What can AI do for me?” It’s “How can I get AI to help me ask better questions?” Try it this week: Before you ask any AI to make or write something, pause and ask another one: “How would you ask this?” That one sentence could double your output without a single new tool. Keep your questions smarter than your answers, — Andrew “Prompt Whisperer” Molloy P.S. If you just realized you've been asking AI the wrong questions for months, you're not alone. That's exactly why we built the Prompt Lab inside the Automation Wizard program — pre-tested templates, system prompts, and meta-prompting frameworks you can swipe and customize in minutes. No guesswork. No fluff. Just the exact structure that turns "meh" outputs into "wait, how did you do that?" Learn more here.
|
Automation Wizard/Soloneer - Exploring how we can use the latest technologies such as AI, automation tools like n8n, and 3D printing to make our lives easier and better.
I once spent £300 on an Iron Man figure because I "earned it." This was my productivity reward system. Hit a revenue milestone, buy a collectible. Complete a major project, add to the collection. Sounds smart, right? Gamify your work. Create incentives. Make the grind feel like a game. Here's what actually happened: Month 1: Hit the milestone, bought the figure, felt genuinely motivated. Month 2: Missed the milestone, bought the figure anyway "for future motivation." Month 3: Tight month...
I run about 2.5–3.5 people's worth of work. Not because I'm exceptional at productivity. Because the automations are good enough that I keep saying yes to things a normal human capacity wouldn't allow. For the past couple of years, I've been the "automation guy" who makes everything look effortless. Client newsletters, social media pipelines, system builds, strategy calls. And I genuinely love it. But behind all that efficiency is a human who still has to think, decide, and solve problems....
I need to tell you something a bit embarrassing. A few days ago, this newsletter sent an issue that was broken. The formatting was wrong. Subject lines leaked into the body. It looked like someone had pasted a template scaffold and hit send without reading it. It was me. I'm the someone. The Automation Wizard, automation-wizard-ing himself into a mess. So it felt like the right moment to share the test I use to make sure AI-assisted content is actually worth sending. It's one question: Remove...