7 Questioning Techniques to Double AI Answer Quality, More Useful Than Switching Tools

Many people think AI is not good enough, but the problem lies not in the tool, but in "how to ask". These 7 questioning techniques I use every day, and after learning them, you'll find that the same AI is like a completely different, stronger model.

I often tell people: instead of constantly switching AI tools, it's better to learn how to ask questions first. The same ChatGPT will yield vastly different results for those who know how to ask and those who don't. In this article, I'll share the 7 questioning techniques I use every day, which don't require any technical background, and after learning them, you'll feel like the AI in your hands has suddenly become smarter.

1. Give it a "role"

Add a sentence at the beginning, "You are a seasoned (某 profession)," and the level of professionalism in the response will be immediately different. If you want legal advice, say "You are a practicing lawyer," and if you want a marketing copy, say "You are a seasoned copywriter." The AI will respond in the tone and with the expertise of that role. This trick is the simplest and most effective.

2. Explain "who it's for and what it's for"

Don't just say "help me write an article." Say "help me write an article for new mothers about complementary foods, with the goal of making them less anxious." The more context you provide, the more accurate the AI's response will be. Treat it like a new colleague who knows nothing about your task, and explain it in a way that you would to them.

3. Provide examples (the most underestimated trick)

Give it one or two examples of what you want, and it will produce something in the same style. If you want it to write a title, paste three titles you like and say "write in this style." Examples are a hundred times more useful than a bunch of adjectives.

4. Do one thing at a time

Don't cram "think of a title, write the content, add images, and translate" into one instruction. Break it down, and do one thing at a time. The quality of each task will be much better.

5. If you're not satisfied, "give feedback" instead of restarting

The essence of AI lies in conversation. If the first version is not good, it's normal. What you need to do is tell it what's wrong: "too formal," "delete this paragraph," or "be more specific." Those who use AI well are "talking" their way to good results, rather than expecting to get it right in one go.

6. Ask it to "ask you questions first"

For complex tasks, you can say "before starting, ask me a few questions you need to know." The AI will clarify the requirements and produce something closer to what you want. This trick is unknown to many people, but it's very useful.

7. Ask it to "explain the reason" or "think from a different perspective"

When you want to delve deeper, ask it to "explain why it made this suggestion" or "think from the opposing side again." This will force out more in-depth answers and help you see blind spots.

The final mindset

The techniques behind these tricks are actually the same: speak clearly. AI can't read minds, and the more complete and specific the information you provide, the more it can help you. This is a skill that gets stronger with practice, and after using it a few times, you'll get the hang of it.

Instead of worrying about "which AI is the strongest," try these 7 techniques with the tools you have in hand today, and you'll find that they're actually more useful than you think. For further reading: Common pitfalls for AI beginners and Advanced ChatGPT usage.

In a nutshell: Instead of constantly switching AI tools, it's better to learn how to ask questions first — give roles, provide context, provide examples, do one thing at a time, and use feedback to chat your way to good results. These 7 techniques are more useful than switching tools.

Sources

Compiled from the editorial team's actual usage experience.

Frequently Asked Questions

How can I get better answers from AI?

Give it a role, clarify who the response is for and its purpose, provide examples, focus on one task at a time, and if you're not satisfied, give feedback and adjust. For complex tasks, let the AI ask you questions first.

Why do I feel that AI is not good enough?

Most of the time, it's not a problem with the tool, but with the questioning approach. The same AI can produce vastly different results for those who know how to ask and those who don't. Practicing questioning techniques is more useful than switching tools.

Does providing examples to AI really work?

Yes, it's extremely useful and often underestimated. Giving one or two examples of what you want can help the AI produce results in the same style, which is much more effective than using a bunch of descriptive words.

Do I need a technical background to master questioning techniques?

Not at all. The core of these techniques is to communicate clearly and provide sufficient context, which is a skill that anyone can develop and improve with practice.

繁體中文版 →