How to Use AI When You Don't Even Know What to Ask (Start with These 4 Questions)
When you're staring at a blank prompt with no idea what to type, these questions can get you unstuck.
You know the feeling.
You sit in front of Claude or ChatGPT knowing you need to learn something, but you have no idea what questions to ask.
So you type something vague like
"Tell me about data science" and get overwhelmed by information that doesn't connect to your actual situation.
The real problem isn't that AI gives bad answers. It's that when you don't know what you don't know, asking for explanations is the wrong approach entirely.
What you need instead are questions that help you figure out what you should be asking in the first place.
The "I Don't Know What to Ask" Problem
When you're starting from zero knowledge, traditional question-asking doesn't work.
You end up with generic overviews that don't connect to your actual work. AI dutifully explains concepts, but you're left wondering "Okay, but what does this mean for me?"
The solution isn't better explanations.
It's better questions - questions designed specifically for when you don't know what you don't know.
But sometimes you're so lost that even forming better questions feels impossible.
When that happens, start here.
The "I'm Completely Lost" Emergency Kit
When you have zero idea what to ask, use these three emergency questions in order:
Step 1:
"I want to [your goal] but I don't know what I don't know. What questions should someone in my position be asking?"Step 2:
"Of those questions, which one would give me the biggest immediate improvement in my work?"Step 3:
"What would I need to know to answer that question properly?"This 3-step sequence can transform "I'm lost" into "I have a clear next step." Use it whenever you're overwhelmed or don't know where to begin.
Once you have direction from the emergency kit, these next 4 question types help you explore more strategically.
4 Types of Questions That Work When You're Starting from Scratch
1. Questions that connect new topics to what you already do
Instead of asking AI to explain concepts in isolation, ask it to build bridges from your current reality.
Don't ask:
"Explain machine learning to me."Ask this:
"I analyze sales data in Excel and create monthly reports. I keep hearing machine learning could improve my analysis. What would I need to learn to get from where I am now to actually using ML in my work?"The difference? You get a custom roadmap instead of generic information.
When I used this approach to understand data visualization better, AI didn't just list chart types. It looked at what I was already doing (basic Tableau charts) and showed me the specific steps to improve and integrate interactive dashboards. In general, this is much more useful than a generic "data viz 101" explanation.
Try adding this to any learning question:
"Given that I currently do [describe your work], what would I need to learn to [your goal]?"This type of question gives you direction. But direction without awareness of pitfalls can lead you astray.
2. Questions that reveal what usually trips people up
Sometimes, the biggest learning breakthroughs come from discovering what you were about to do wrong.
Before diving into any new topic, ask about the common mistakes:
"What assumptions am I likely making about [topic] that could limit my understanding? What do people typically get wrong when they're starting out?"This can save you a lot of headaches.
The point isn't to become an expert immediately. It's to learn what you don't know so you can ask better questions as you go.
Knowing the pitfalls helps, but one perspective can still leave you with blind spots.
3. Questions that give you multiple viewpoints
When you don't know what you don't know, one perspective isn't enough.
Instead of asking
"What are dashboard design best practices?" Try
"What would a UX designer, a data analyst, and a business stakeholder each say are the most important aspects of dashboard design?"Each perspective can reveal something different.
The UX person talks about cognitive load.
The analyst focuses on data accuracy.
The stakeholder cares about decision-making speed.
Suddenly you're not just making prettier charts - you're thinking about how different stakeholders actually use information.
Multiple viewpoints help you understand the full scope of a topic. But all this exploration only helps if you can turn it into action.
4. Questions that create realistic next steps
All the insights in the world won't help if you can't figure out what to do first.
Ask AI to factor in your actual situation:
"I have [time constraint] and currently know [your skills]. Create a realistic learning plan for [your goal]. What should I focus on first, and what can wait?"The key word is "realistic." AI can factor in what you already know and give you a progression that actually makes sense for your situation.
No more wondering if you're learning things in the right order or if you're missing crucial foundations.
Why This Changes How You Learn
When you don't know what to ask, the solution isn't better answers - it's better questions.
The emergency kit gives you a starting point when you're completely stuck.
The four question types give you depth once you know where you're headed.
Together, they can turn AI from an information dispenser into a thinking partner.
Instead of drowning in generic explanations, you get guidance that connects to your real situation. You're not trying to learn everything - you're trying to learn what matters for where you are right now.
The people who advance in data work aren't the ones with perfect knowledge. They're the ones who get comfortable with uncertainty and use it to ask better questions.
What will you try first?
Chat soon,
Donabel
The gap between "I don't know what to ask" and "I know exactly what I need to learn" is just a few good questions away.