Teknik “Right-Question” Bantu Pengguna Perbaiki Prompt ChatGPT

Refining ChatGPT Prompts with Critique

Users frustrated by long or off-target ChatGPT responses often find the issue lies in the prompt itself rather than the AI’s processing power. According to KompasTekno, a technique known as the “right-question” approach allows users to force the chatbot to evaluate the quality of a prompt before generating a full answer.

This method requires users to add a specific instruction to their initial query. By asking the AI to critique the question’s effectiveness for a specific goal, users can identify if their framing is too broad or missing key context.

Applying the Right-Question Technique

To implement this strategy, users should append a directive to their request. As reported by PCWorld, an effective prompt structure looks like this: “Before you answer my request, critique the question itself. Is this the right question to be asking to achieve [my goal]? If not, suggest a better question and explain why.”

When this instruction is used, the AI shifts from simply generating text to performing an internal audit of the user’s intent. If the AI determines the original question is misaligned with the stated goal, it will provide alternative phrasing and explain the reasoning behind those changes.

When to Use Prompt Auditing

This technique is most useful for complex tasks that involve deep research, product comparisons, or exploring multifaceted phenomena. Because these tasks often require specific data points, the AI’s feedback helps narrow the scope before the actual search begins.

However, this process is not always necessary. For highly specific or routine tasks, adding a critique instruction can waste time and computational tokens. The method is best reserved for situations where the user is uncertain about the best angle for a query or needs help defining the parameters of a large research project.

Commonly Asked Questions About Prompt Engineering

Does this technique work for all AI models?

The "right-question" technique uses the conversational evaluation capabilities present in most large language models like ChatGPT.

Can I ask the AI to refine its suggested questions further?

Yes. Users can repeat the critique instruction after receiving a revised question from the AI. This iterative loop allows for multiple rounds of refinement until the prompt is optimized for the intended research outcome.

Is this method effective for simple, factual queries?

No. For direct questions—such as asking for a specific date or a simple definition—adding a critique instruction is unnecessary and may result in an overly verbose or inefficient conversation flow.

The effectiveness of this approach relies on the user’s ability to clearly define their end goal within the initial prompt. If the goal is vague, the AI’s critique will likely reflect that ambiguity, necessitating further clarification from the user.