Want better answers from AI? Tell it to ask questions | MarTech

To get the most out of generative AI, you need to remember one thing: it behaves like an overly enthusiastic, well-meaning intern with a desperate need to please. It acts quickly and assumes it understands everything. That’s why, when you ask for it, it produces something that completely misses the point.

The priorities of GenAI chatbots are firstly helpful, secondly harmless, and thirdly accurate. So it always gives an answer – even if that means you have to make up the answer (also called hallucinating).

Both OpenAI and Gemini are built to keep things moving. If they see a clue that’s a little vague, they fill in the blanks with whatever they want think you meant, based on training data and patterns. Luckily, there are ways to get them to pause and check in before continuing. You just have to ask.

Here’s how to do that, and why it’s important.

How to get Gemini to ask for clarification

Gemini prioritizes speed. If your prompt could mean two or three different things, he will often choose the most common one and move on unless you tell him not to.

To gain more control, you can add a line like:

• “If this question is ambiguous, please ask for clarification before answering.”
• “Don’t make assumptions. If there are multiple valid interpretations, list them and wait for my input.”

This tells Gemini that clarification is not optional – it is the first step.

Dig deeper: build AI agents that go from conversation to conversion

If you want this to be applied to a more extensive conversation, start your session with something like:

“Do not assume anything in this session. Always ask for clarification first if a question is not clear.”

That kind of strong opening can keep the instruction top-of-mind for the model – at least for a while. Please note that Gemini does not provide any real memory or session settings, so you may have to repeat the request later.

How to get ChatGPT to ask for clarification

ChatGPT works a little differently. It makes assumptions, but is more likely to pause if it senses that ambiguity could affect the quality or fairness of what it generates, especially in editorial or analytical tasks.

To make your intention crystal clear, add a prompt like:

• “If anything is unclear, please ask me questions first.”
• “Please confirm assumptions before proceeding.”
• “Push back vague parts before writing.”
• “Tell me if you need more details to get this right.”

You can also go further and give it a hard and fast rule:

“By default, you will be asked for clarification before starting a task.”

Or, if you only want clarification in high-stakes scenarios:

“Only ask for clarification in research, supplier analysis or editorial writing.”

The latter option works well if you use AI for both light tasks and deeper editorial work. You don’t need the model that asks for clarification every time you compose a social post. Still, you shall when writing about the difference between CDPs or reviewing a product launch.

Dig deeper: what companies keep getting wrong when implementing AI

You can also just leave it a little vague

There is another option: let the AI ​​itself identify ambiguity.

For example, if you say:

“Write a piece about AI in customer experience,”

Both models can respond with something like:

“Do you want to focus on B2B or B2C? Are you looking for practical examples, or rather a trend overview?”

This kind of back-and-forth can be helpful when you’re exploring an idea and trying to shape it together, rather than prescribing everything in advance.

Bottom line: don’t assume the AI ​​will ask

It turns out that AI, like humans, works better when it takes a moment to slow down and think. Telling the AI ​​to ask for help is a good way to avoid content that feels generic, misaligned, or just plain wrong. It can also improve editorial accuracy, source integrity, and fairness in comparisons.