Someone handed me a 6-prompt growth system

If you’ve been on Threads, Instagram, Twitter, anywhere on the internet lately, you’ve seen the posts. “I talked to an Instagram growth strategist and put everything she asked me into Claude. Here are the six questions you should be asking.” Six prompts, copy-paste, watch your account grow.

Honestly? They’re not bad. But they’re not always right.

Let’s be clear about what these posts are. They’re engagement bait, every single one of them. And they’re also teaching you something useful: a way of thinking, a way of talking to these tools, that you can borrow and make your own. So I don’t hate them. I screenshot them constantly. I just use them as reference for the kinds of things I should be asking my tools, instead of running them word for word.

Because that’s the actual value here. With these tools, we don’t know what we don’t know. A prompt post is a free reminder of question types you haven’t thought to ask yet. Take the idea. Skip the script.

So in that spirit, here are a few of the things I actually ask my tools. Fair warning: all of these require connecting your tool of choice, whether that’s Claude or ChatGPT, to your own data through MCP (short for Model Context Protocol, the standard plug that lets an AI read your other apps). I use Claude for most of them.

“Go through all of my Granola meeting notes from the past month. How can I improve my communication? Are there tasks where we dropped the ball? Are we getting consistent feedback from clients, positive or negative? How can I improve as a leader, and how can my team be doing better? When my team is on calls and I’m not there, what am I missing? Does their performance dip when I’m not present?”

That’s close-to-real-time feedback on your own leadership, pulled from what actually happened in your meetings. Most people get that once a year in a review, if ever.

“Go through my Oura ring data and my food log from the past month. What trends are you seeing? Are certain meals, or meal times, affecting my energy, my sleep, my workouts? What blind spots do I have? Where can I be doing better?”

Same move, personal side. The data was already sitting there. Nobody was reading it.

I built a bot, connected to Telegram, that logs all of my meals every day. For me that’s so much easier than a MyFitnessPal-style app: I text it what I ate, it already knows my regular restaurant orders and their base calorie counts, and it coaches me in real time. “You’re behind your protein target so far today, you might want to grab your protein shake.”

And I still work with a human nutritionist. I’m a big believer in hiring humans, especially expert humans. I export the data and she does the interpretation, and hers is the judgment that wins. She’s the one who said “your protein goal is actually too high for what you eat and your goals, drop it down.” Then I take her feedback and put it back into my agent, so the bot keeps learning and keeps coaching me between our check-ins. The AI handles the logging and the pattern-spotting. The expert handles the thinking.

A business partner of mine wants to expand into a market we haven’t worked in before. So I gave Manus (an AI research agent) the brief and the customer overview, and asked: where’s the area of opportunity? What are our blind spots? What do we need to think about? What’s the revenue potential, and the profit potential?

I love this use because it cites every source, so I don’t feel like it’s guessing or hallucinating. For research tasks like this I reach for Manus or Perplexity over Claude or ChatGPT. They’re simply better research agents, and I trust what comes back.

Notice what every one of those has in common. None of them came from a carousel. They all point the tool at data that’s mine: my meetings, my health, my clients, my market. The six-prompt posts hand everyone the same six questions, and that’s exactly why they produce the same answers for everyone.

Borrow the thinking. Then aim it at your own data.

The prompt was never the valuable part. Your context is.

Cheers,

Audrey