Imagine you’re working with an enthusiastic intern. You give them instructions and they get on with their task. When they return to present their work, it’s not right. They missed the mark. They didn’t follow instructions. Ambiguity crept in and now the output isn’t what you wanted. Something got lost in translation.
What would you do to correct their work? You wouldn’t just say “do it again” without helping them out. You wouldn’t send them away without explaining why. Apply this same thinking to ChatGPT.
When you get suboptimal output from ChatGPT, the easy thing to do is hit “regenerate.” But actually, that’s a bad idea. Here’s why you should stay away from that tempting button, and here’s what to do instead.
Get better results from ChatGPT: how to re-prompt for best results
Practice infinite patience
Rather than treating ChatGPT like a kitchen appliance you can get frustrated with, treat it like a member of your team. Commit to re-prompting until you get it right. Don’t give up and close your laptop.
Stick with the tool until the output is optimal. Be patient with the process. Rushing only leads to more mistakes.
Assume the problem is you
Your prompts might seem crystal clear to you, but ChatGPT needs specific direction. When you get odd results, assume you did something wrong. Look at your prompt with that in mind. Did you leave room for interpretation? Were your requirements fuzzy? How could you have done better?
Tighten up your instructions before blaming the AI. Team up to succeed together. Take ownership of the clarity in your requests.
Play detective
Once you’ve analyzed your original input, study the output. Spot where things went wrong. Was it the tone? The structure? The length? Mark these issues and address them one by one in your follow-up prompt. Go back through your original instructions and spot the ambiguous terms. Avoid anything that requires extra thinking.
Tell ChatGPT exactly what missed the mark and guide it toward your target. Each edit brings you closer to perfect output.
Be specific
Don’t leave the changes to chance. List your exact requirements. Call out exactly what’s wrong, and share examples of what you want. The more precise your feedback, the better your results. If you don’t have time to type, hit record and voice note the edits. The robot doesn’t mind.
Tell ChatGPT exactly which parts need work and what good looks like to you. Small details make big differences in your results.
Stay focused on quality
Keep your eyes on the prize, which is high-quality output. Don’t settle for good enough when you can get exactly what you need with the right prompts and re-prompts. Track what works and repeat those patterns. Build a mental library of successful approaches, and store your prompts somewhere you easily re-use. Go back through previous chats where the output was correct, and find the phrases that generated better results.
Excellence comes from attention to detail and knowing what you want. Your standards matter, so communicate them clearly.
Stop regenerating, start communicating: how to prompt ChatGPT
Just like managing a team member, working with AI takes skill, compassion and clear instructions. Stay away from the regenerate button. Instead, build your prompt expertise. Learn to spot weak instructions. Give better feedback. Guide the AI with patience and precision. Your results will improve dramatically when you treat each interaction as a chance to refine your approach.
Every conversation with ChatGPT can produce great work as long as you use clear instructions and give thoughtful feedback. Start treating ChatGPT like a teammate who wants to help, and watch your results improve.
Read the full article here
