You've probably already experimented with AI tools like ChatGPT, CoPilot or Google Gemini.
And maybe the first time you asked something like:
- "Write a 500-word article about the benefits of solar panels."
- "Give me a list of 10 healthy breakfast recipes."
- "Create a marketing plan for a new gym."
Did you always get exactly the output you wanted then? Probably not. And that's to be expected.
Could it possibly be done differently?
Be patient with your prompts
"Hey ChatGPT, write a blog post on climate change"
This is how many people use AI. They throw in a short question, with little to no context, and then expect a magical end result bursting with creativity and originality.
And when that result doesn't come, they blame AI. I don't think that's entirely fair.
Couldn't we treat AI a little more as an equal, by giving it the same attention as a human interlocutor? Shouldn't we be a little more patient with AI?
A real conversation
Imagine this: you have a new colleague. You ask for a report on a project you just completed. Would the instruction be enough for you then: "Just make a report on that project from last week."
No right?
You would also start with a real colleague by providing explanation and context, asking and answering questions, aligning expectations. Anything necessary to create clarification: going into depth to get to the core.
It's the same with AI. It's smart, fast, and does exactly what you ask, but it doesn't yet know you and your needs. When you first start using ChatGPT, it doesn't yet know what your goals are, what requirements you have, and what the end result should be.
And that makes sense, too. A flawed input is a flawed output.
Instead of immediately starting to give orders, it's better to first take some time to tell what you need, what the background is, what the goals are and most importantly, what you care about the end result.
You can't do that in 1 or 2 sentences. That takes a good conversation.
A prompt is not the same as a query
A search engine like Google is especially fast. Type a few words and you get a list of results.
AI models are different. It wants to know more about you. It asks questions back. Why are you looking for this? What do you want with it? Not a quick list, but a comprehensive conversation.
AI is about engaging in a dialogue in which you and the language model can come to insights together. AI wants to understand you as best it can. It learns from what you say and continuously builds on that.
By taking the time to explain your thoughts, goals and expectations, you give an AI model the opportunity to generate tailored, relevant and valuable output.
So AI is much more than just a simple provider of information. It is like a close friend or associate that you get to know better and better.
The more often you meet, the better you will understand each other.
Quality takes time
Generating qualitative output with AI takes time. And not because the language model is slow, but because you as the user need to take the time to explain as best you can what your goal is, what you care about and what you really want to achieve.
Just like in a real conversation, provide enough context, take time to answer questions and correct inaccuracies where necessary to discuss a complicated situation thoughtfully.
AI is not clairvoyant
AI is smart, fast and efficient. But certainly not clairvoyant.
Therefore, it can often be helpful to clarify what you don't want with "negative prompts.
- "Prefer no jargon" - If you want to keep it simple
- "Avoid overly long sentences" - For a more active writing style
- "No technical details" - If you want to keep it accessible to a wider audience
If necessary, clarify these instructions with concrete examples from the generated output.
By bringing up what you don't like in this way, you help AI find the right direction faster.
Think of it as drawing out a map, including a list of crucial places to avoid.
Get questioned
Usually you ask something of AI. But turn the tables. Let AI question you properly.
Try something like: "Play my critical supervisor for a moment. What five things are you still missing from my proposal?"
This often delivers more than you think, because:
- You see your plan through different glasses
- Gaps in your story become visible
- You discover new perspectives
Then when your AI tool makes suggestions like, "Have you thought about...?", you know: Oh yes, I need to do something about that.
Learn from mistakes
Perfect output all at once? That's rare. And I don't think you should want this either. Adjusting is okay. Every "mistake" by the AI is a new opportunity for you to fine-tune.
For example, during your chat session with ChatGPT, say, "This part is good, but can you edit that bit after?"
This is how you and your AI tool get to know each other better and better. If you invest a little more time in this at the beginning of the conversation, you will see that in the long run your output will become more and more in line with your expectations.
Conclusion
Interaction with AI is much more like human interaction than you might think.
- Experience teaches you what works and what doesn't.
- Take the opportunity to make attentive adjustments when necessary.
- Add details, correct with appropriate prompts and clarify your intentions.
- With each interaction, you learn more about the capabilities and limitations of AI.
- At the same time, AI is getting to know your personal preferences and style better and better.
This process simply requires patience and attention.
Sometimes you have to take a step back to take two steps forward. You experiment, adjust, and try again. The goal is not perfection in one attempt, but gradual improvement. With each iteration, you get closer to your desired result. Build a bridge, so to speak, between your ideas and the capabilities of AI.
Eventually you then reach a point where the output not only meets your expectations, but may even exceed them.
Qualitative output with AI does not come your way.
It takes time to grow. Step by step.
Just like with real people.
Related articles
Nederlands