Yes, this is exactly how to get the best out of ChatGPT. You just have to have a conversation with it and ask it questions. Refine it. Clarify your own thinking. But it takes time to do that. It isn't just a prompt it and go. I don't even like using the word prompt.
I think the real issue is that people think it's this magic wand that you just tell it to do something, and it gives you exactly what you want without much effort. Which, to be honest, is how it's been advertised in a lot of ways. But to truly get the best benefit from it, you have to be willing to put the time in.
Absolutely - definitely not a magic wand and can’t read minds. You need to train it just like you would a new colleague, by showing it how you work, what you want, how you think etc. - so that it can make the most of the knowledge it has and apply it on your situation
Great article. I think the instinctual obstacle is generational, which is (for once!) a good thing for old fogeys like myself. Wrestling with ideas, asking questions, challenging, provoking—these were all taught to me at an early age (as being fun and exciting, believe it or not!), and then supported throughout my schooling.
We also didn’t have quite so many distractions, nor as a result the quick-fix expectations Z and Alpha, especially, have when it comes to demanding results without putting in the work.
AI has to be taught with the expectations you’d have for any process-dependent subject: show me the steps, the thinking, the back and forth work that helped you arrive at this output. No steps—no engagement to show me your (or you’re!) thinking and logical understanding—no mark. Business must demand this, too, if at the least for legal and reputational self preservation!
That’s really interesting to hear - where did you grow up?
I’m based in Romania, born and raised here. Our schools were definitely more traditional. We weren’t really encouraged to question things or speak up much. I was lucky to have some amazing teachers along the way, but for my parents’ generation, that kind of open inquiry just wasn’t part of the system. Makes sense though, considering we were a communist country until 1989, the legacy of that is still visible in how education approaches authority and conformity. So yeah, culture and history has a lot to do with this.
That said, I do think things are shifting now and there’s more openness compared to how things used to be before.
And yes, totally agree with you on the distraction front. We’re all developing attention deficit tendencies just from how tech and social media have evolved. Which is exactly why I believe AI should be introduced in schools, not banned. Kids will use it anyway. The responsibility then is to teach them to use it well.
interesting question, I think the benefits coming from this recursive prompting method are super broad - cause 1) they allow me to learn and explore a lot of things I wouldn’t be confident enough to approach and that would take me a very long time to learn on my own, and 2) it speeds up how fast I can implement all the strategies, which results in actual outputs.
It’s 11pm where I am, so this what I can think of right now, haha.
& let me know if you try / tried this approach, super curious how it’s been for you!
Benta, you just articulated the exact thing I was trying to get at, that recursive prompting isn’t just a technique, it’s a way of thinking that rewards curiosity and depth - not operating on autocomplete.
Absolutely. The process is where growth happens where we stumble, stretch, and become. The destination may be the goal, but the journey shapes the soul.
In both life and my writing, especially in B2B health, I’ve learned that showing up for the process is the real work. Thank you for this grounding reminder.
I know there is a lot more to this blog than the graph, but it really caught my eye.
It’s fascinating to see how emerging economies like India, Nigeria, and the UAE are leading in both AI adoption and trust at work.
This really highlights how necessity and openness to innovation can accelerate tech integration faster than regulation-heavy or risk-averse environments.
On the flip side, the gap in trust across some Western countries, despite moderate usage, suggests a real need for clearer communication about AI’s role, limitations, and governance.
Culture, policy, and digital maturity are clearly shaping very different AI trajectories worldwide.
Yes! In many emerging economies, there’s less institutional drag and often a stronger drive to leapfrog when digital tools can fill structural gaps or open doors that didn’t exist before. And beyond infrastructure, there’s also more hunger to catch up, compete, succeed. That urgency also makes the adoption to happen faster.
I was also really surprised to see Romania (my country) ranking right alongside the US.
Had a really good friend that was Romania when i lived in Germany. He liked the work in Germany but ALWAYS talked about the food in Romania. The best bread. The best meat. The best vegetables from the farms… 🤣 It got to be a joke after a while.
This is beautifully said. I've been calling it “warming the model”, that process before the prompt, where clarity forms through the conversation itself.
For me, the best prompt is rarely the first one. It's the 9th, or 19th, or 49th. Because it's not really a prompt, it's a condensation of everything I’ve just understood in dialogue with the model.
You’re also pointing to something deeper: how our results-obsessed culture trains us to skip the part where meaning gets made. That urgency to ship, go viral, or “just get it done” shows up in how we prompt too, and often leaves us with output that looks good but lacks alignment.
I’m working on a piece about this shift from AI as task-doer to AI as thinking partner. This essay captures that transition so well.
We’re literally warming the model, I love that framing. What I’ve noticed with this approach is that it makes traditional prompting techniques less important than usual. But it also makes the context you bring into the conversation absolutely essential.
It’s such a simple and accessible thing, but at the same time so powerful. I use it every day. Like today, I wasn’t even tackling a new task, just working on some Facebook ad angles for Playsense. And as you said, I was “warming it up” by going deeper into the problems the target group faces, how they try to solve them now, what’s still missing, what they talk about online etc. That process helped narrow things down before I even started generating copy - which not only made the ads better, but also gave me new insights.
Really appreciated how clearly you broke this down, Daria. That three-part framework (expertise, prompting technique, and personal context) gets to the heart of why better prompts alone aren’t enough. I’ve found myself doing a version of this often, and your examples captured the process really well.
The way you explained recursive prompting especially hit home. At work, when I’m tackling something I don’t fully understand, my first instinct is to find the subject matter expert, set up a one-on-one, ask a bunch of questions, take notes, and try to apply what I learn to the problem at hand. It’s how we learn on the job.
Using AI this way feels like an extension of that same approach with a few workplace perks. The “expert” never has calendar conflicts, doesn’t reschedule last minute, and is somehow available at 11PM when I’m trying to meet a deadline. And instead of just absorbing the insight, I can apply it immediately in the same session :)
I love the comparison to how we naturally learn on the job, and how recursive prompting is exactly that… Didn't think about it that way, but it makes perfect sense!!
And you’re right, the magic isn’t just in getting answers, but applying them right away while the insight’s still fresh. That's why this prompting technique works so well, cause it's not something you just do for the future, it helps you with the task you need to do right now. And it's super easy to act on it.
Haha yesss! GPT actually told me the name after I dumped my messy draft. Recursive prompting inception 😂 Then I dug into it more and realized, oh… this is what I’ve been doing.
And omg yes, Canva is the best! I’ve been using it for 5+ years. Their AI + integration game has leveled up a lot lately. Enjoy the rabbit hole!
You’re capturing the infinite prompting concept so well, Daria! This really feels like a necessary step in learning to become the true agent of our own thinking, instead of outsourcing it entirely to AI.
I also love how clearly you broke down the actionable steps, they’re easy to follow and incredibly grounding. Thank you for putting this into words so powerfully!
Thanks so much, Jenny, that means a lot! 🙏 I really appreciate you picking up on the core idea of claiming our role as the thinker, not just the task assigner.
And yeeees, it does look like infinite prompting at first glance, haha, as the loop and conversation goes on and on, but it's different than @Michael Simmon's "infinite prompting" framework.
Only part I disagree with is when you stated “I’m not an expert”,
Haha I beg to differ
hahaha Joel, I was winging it at first and still am a lot of the time. But I really appreciate the vote of confidence!!
A really detailed and informative analysis.
“In a world where every knowledge worker worries about being replaced by AI, becoming someone who knows how to use it properly is a superpower”.
This is powerful!
Appreciate that! And yep, knowing how to use it well is what separates shortcuts from leverage
Such a powerful breakdown. Once you get used to questioning before producing, you start approaching non-AI tasks the same way.
Exactly, it’s all in the questioning and curiosity. That’s the real recipe for staying sharp and never stopping the learning.
Yes, this is exactly how to get the best out of ChatGPT. You just have to have a conversation with it and ask it questions. Refine it. Clarify your own thinking. But it takes time to do that. It isn't just a prompt it and go. I don't even like using the word prompt.
yesss, the real magic happens in the conversation. It does take a bit more time, but the payoff is way higher.
I think the real issue is that people think it's this magic wand that you just tell it to do something, and it gives you exactly what you want without much effort. Which, to be honest, is how it's been advertised in a lot of ways. But to truly get the best benefit from it, you have to be willing to put the time in.
Absolutely - definitely not a magic wand and can’t read minds. You need to train it just like you would a new colleague, by showing it how you work, what you want, how you think etc. - so that it can make the most of the knowledge it has and apply it on your situation
Great article. I think the instinctual obstacle is generational, which is (for once!) a good thing for old fogeys like myself. Wrestling with ideas, asking questions, challenging, provoking—these were all taught to me at an early age (as being fun and exciting, believe it or not!), and then supported throughout my schooling.
We also didn’t have quite so many distractions, nor as a result the quick-fix expectations Z and Alpha, especially, have when it comes to demanding results without putting in the work.
AI has to be taught with the expectations you’d have for any process-dependent subject: show me the steps, the thinking, the back and forth work that helped you arrive at this output. No steps—no engagement to show me your (or you’re!) thinking and logical understanding—no mark. Business must demand this, too, if at the least for legal and reputational self preservation!
That’s really interesting to hear - where did you grow up?
I’m based in Romania, born and raised here. Our schools were definitely more traditional. We weren’t really encouraged to question things or speak up much. I was lucky to have some amazing teachers along the way, but for my parents’ generation, that kind of open inquiry just wasn’t part of the system. Makes sense though, considering we were a communist country until 1989, the legacy of that is still visible in how education approaches authority and conformity. So yeah, culture and history has a lot to do with this.
That said, I do think things are shifting now and there’s more openness compared to how things used to be before.
And yes, totally agree with you on the distraction front. We’re all developing attention deficit tendencies just from how tech and social media have evolved. Which is exactly why I believe AI should be introduced in schools, not banned. Kids will use it anyway. The responsibility then is to teach them to use it well.
Thanks for sparking this conversation!
Daria - did you receive an email from me on Fearless Investor community?
Hey Neil, yes, I did!
Another great article - how are you experiencing the benefit in your business?
interesting question, I think the benefits coming from this recursive prompting method are super broad - cause 1) they allow me to learn and explore a lot of things I wouldn’t be confident enough to approach and that would take me a very long time to learn on my own, and 2) it speeds up how fast I can implement all the strategies, which results in actual outputs.
It’s 11pm where I am, so this what I can think of right now, haha.
& let me know if you try / tried this approach, super curious how it’s been for you!
There’s something beautifully human about returning to a question again and again, each pass deepening the texture of the answer.
Recursive prompting mirrors the way we think when we’re truly curious circling, refining, discovering nuance in places we once thought were flat.
Your piece captures that quiet patience so well. It’s a reminder that depth isn’t rushed; it’s invited.
Thank you for this. It feels like an open window for anyone willing to explore a little further.
Benta, you just articulated the exact thing I was trying to get at, that recursive prompting isn’t just a technique, it’s a way of thinking that rewards curiosity and depth - not operating on autocomplete.
Absolutely. The process is where growth happens where we stumble, stretch, and become. The destination may be the goal, but the journey shapes the soul.
In both life and my writing, especially in B2B health, I’ve learned that showing up for the process is the real work. Thank you for this grounding reminder.
I know there is a lot more to this blog than the graph, but it really caught my eye.
It’s fascinating to see how emerging economies like India, Nigeria, and the UAE are leading in both AI adoption and trust at work.
This really highlights how necessity and openness to innovation can accelerate tech integration faster than regulation-heavy or risk-averse environments.
On the flip side, the gap in trust across some Western countries, despite moderate usage, suggests a real need for clearer communication about AI’s role, limitations, and governance.
Culture, policy, and digital maturity are clearly shaping very different AI trajectories worldwide.
Yes! In many emerging economies, there’s less institutional drag and often a stronger drive to leapfrog when digital tools can fill structural gaps or open doors that didn’t exist before. And beyond infrastructure, there’s also more hunger to catch up, compete, succeed. That urgency also makes the adoption to happen faster.
I was also really surprised to see Romania (my country) ranking right alongside the US.
Had a really good friend that was Romania when i lived in Germany. He liked the work in Germany but ALWAYS talked about the food in Romania. The best bread. The best meat. The best vegetables from the farms… 🤣 It got to be a joke after a while.
hahah, love that. One thing I also learned while living abroad is how clearly you start to see and miss the best parts of your home country
This is beautifully said. I've been calling it “warming the model”, that process before the prompt, where clarity forms through the conversation itself.
For me, the best prompt is rarely the first one. It's the 9th, or 19th, or 49th. Because it's not really a prompt, it's a condensation of everything I’ve just understood in dialogue with the model.
You’re also pointing to something deeper: how our results-obsessed culture trains us to skip the part where meaning gets made. That urgency to ship, go viral, or “just get it done” shows up in how we prompt too, and often leaves us with output that looks good but lacks alignment.
I’m working on a piece about this shift from AI as task-doer to AI as thinking partner. This essay captures that transition so well.
We’re literally warming the model, I love that framing. What I’ve noticed with this approach is that it makes traditional prompting techniques less important than usual. But it also makes the context you bring into the conversation absolutely essential.
It’s such a simple and accessible thing, but at the same time so powerful. I use it every day. Like today, I wasn’t even tackling a new task, just working on some Facebook ad angles for Playsense. And as you said, I was “warming it up” by going deeper into the problems the target group faces, how they try to solve them now, what’s still missing, what they talk about online etc. That process helped narrow things down before I even started generating copy - which not only made the ads better, but also gave me new insights.
& can’t wait to read your piece!!
Thank you Daria, I feel we see it very similarly. Will post it tomorrow hopefully.
"The robots aren't taking over. They're waiting for you to lead." That's the one, right here. 💛
Really appreciated how clearly you broke this down, Daria. That three-part framework (expertise, prompting technique, and personal context) gets to the heart of why better prompts alone aren’t enough. I’ve found myself doing a version of this often, and your examples captured the process really well.
The way you explained recursive prompting especially hit home. At work, when I’m tackling something I don’t fully understand, my first instinct is to find the subject matter expert, set up a one-on-one, ask a bunch of questions, take notes, and try to apply what I learn to the problem at hand. It’s how we learn on the job.
Using AI this way feels like an extension of that same approach with a few workplace perks. The “expert” never has calendar conflicts, doesn’t reschedule last minute, and is somehow available at 11PM when I’m trying to meet a deadline. And instead of just absorbing the insight, I can apply it immediately in the same session :)
I love the comparison to how we naturally learn on the job, and how recursive prompting is exactly that… Didn't think about it that way, but it makes perfect sense!!
And you’re right, the magic isn’t just in getting answers, but applying them right away while the insight’s still fresh. That's why this prompting technique works so well, cause it's not something you just do for the future, it helps you with the task you need to do right now. And it's super easy to act on it.
Thanks for your thoughtful comment!
I didn’t know it had a name either LOL. #teamrecursiveprompting all the way. I’m currently going through that process to learn how to use canva
Haha yesss! GPT actually told me the name after I dumped my messy draft. Recursive prompting inception 😂 Then I dug into it more and realized, oh… this is what I’ve been doing.
And omg yes, Canva is the best! I’ve been using it for 5+ years. Their AI + integration game has leveled up a lot lately. Enjoy the rabbit hole!
Informative, thoughtful, and well done as always. Thank you for sharing this.
Thank you, Sae, always happy if I bring something useful.
You’re capturing the infinite prompting concept so well, Daria! This really feels like a necessary step in learning to become the true agent of our own thinking, instead of outsourcing it entirely to AI.
I also love how clearly you broke down the actionable steps, they’re easy to follow and incredibly grounding. Thank you for putting this into words so powerfully!
Thanks so much, Jenny, that means a lot! 🙏 I really appreciate you picking up on the core idea of claiming our role as the thinker, not just the task assigner.
And yeeees, it does look like infinite prompting at first glance, haha, as the loop and conversation goes on and on, but it's different than @Michael Simmon's "infinite prompting" framework.
Ahhhhh please excuse my bad spelling 🤦♀️
You are so right about the difference between the recursive and infinite prompting!