Flip the Script: Let AI Interview You ← Recommended
You don’t need to become a prompt engineer. You just need to let AI ask the right questions.
Flip the Script: Let AI Interview You
Most people treat AI like a search engine. Type a question, get an answer.
The problem? You spend ten minutes crafting the “perfect prompt.” You still get generic garbage. You try again with different wording. More garbage. You give up and do it manually.
The thing is, the issue isn’t your prompting skills. It’s that AI doesn’t know what it doesn’t know about YOUR situation.
The fix is simple. Stop asking AI questions. Make AI ask you questions instead.
The Problem With Traditional Prompting
When you type “Write me a marketing plan” into ChatGPT, the AI has to make assumptions. What kind of business? What size? What budget? What goals? What’s already been tried?
It doesn’t know any of this. So it produces a generic marketing plan that could apply to any business, which means it applies well to no business.
You might try to fix this by cramming more detail into your prompt. “Write me a marketing plan for a UK-based B2B recruitment agency with 12 staff targeting mid-sized manufacturing companies with a budget of £3,000 per month focusing on LinkedIn and content marketing.”
Better. But you’re now doing the hard work of figuring out what information matters. You’re essentially becoming a prompt engineer, which isn’t what you signed up for.
And even with all that detail, you’ve probably forgotten something important. There’s context in your head that you didn’t think to include because you didn’t know it was relevant.
There’s a better way.
The Interview Flip
Instead of telling AI what you want, tell it what you’re trying to achieve and ask it to interview you.
The prompt is simple:
“I need a marketing plan. Before you write anything, ask me questions until you have everything you need to give me something genuinely useful.”
That’s it.
Now AI becomes the interviewer. You become the expert on your own business. The AI asks about your audience, your goals, your constraints, your past efforts, your competitive landscape. It extracts the context it needs rather than making assumptions.
And here’s what’s remarkable: AI is often better at knowing what questions to ask than you are at knowing what information to provide.
It will ask about things you forgot. It will probe areas you assumed were obvious. It will dig into specifics you would have glossed over.
The result is vastly better output because AI finally has the context it needs.
┌─────────────────────────────────────────────────┐
│ THE FLIP-THE-SCRIPT METHOD │
├─────────────────────────────────────────────────┤
│ │
│ OLD WAY (You → AI) │
│ ┌─────────────┐ ┌─────────────┐ ┌─────────┐ │
│ │ You craft │→ │ AI makes │→ │ Generic │ │
│ │ long prompt │ │ assumptions │ │ output │ │
│ └─────────────┘ └─────────────┘ └─────────┘ │
│ ↓ │
│ Frustration → Rework → More frustration │
│ │
│ ───────────────────────────────────────────── │
│ │
│ NEW WAY (AI → You → AI) │
│ ┌─────────────┐ ┌─────────────┐ ┌─────────┐ │
│ │ Simple goal │→ │ AI asks │→ │ You │ │
│ │ + "ask me" │ │ questions │ │ answer │ │
│ └─────────────┘ └─────────────┘ └────┬────┘ │
│ ↓ │
│ ┌─────────────┐ ┌─────────┐ │
│ │ Tailored │← │ Rich │ │
│ │ output │ │ context │ │
│ └─────────────┘ └─────────┘ │
│ │
├─────────────────────────────────────────────────┤
│ Better context = Better output. Every time. │
└─────────────────────────────────────────────────┘
The Voice Superpower
This method becomes even more powerful when you combine it with voice-to-text tools.
When you type responses to AI’s questions, you tend to be brief. You summarise. You leave things out because typing is effort.
When you talk your responses, you naturally elaborate. You tell stories. You include context you wouldn’t have bothered typing. You explain the nuance.
Tools like Wispr Flow let you speak naturally and have your words transcribed accurately. Instead of typing a two-sentence answer to each of AI’s questions, you can talk for a minute or two, giving rich, detailed responses.
This additional context dramatically improves AI’s output. The difference between a typed “We target mid-sized manufacturing companies” and a spoken “We target mid-sized manufacturing companies, specifically ones in the Midlands with 50 to 200 employees, usually the operations director or MD is our buyer, they’re typically frustrated with high turnover in skilled positions...” is night and day.
You’re not changing what you know. You’re just communicating it more completely because voice removes the friction.
A Real Example
Let me show you how this works in practice.
A recruitment agency owner needs a pricing strategy for a new service. Here’s the traditional approach:
Prompt: “Create a pricing strategy for a new recruitment service targeting UK manufacturing companies.”
Result: A generic pricing framework with common models (retainer, contingency, hybrid), standard percentage ranges, and advice that could apply to any recruitment firm anywhere.
Now the flip-the-script approach:
Prompt: “I need to create a pricing strategy for a new recruitment service. Before you develop anything, interview me until you understand my business, my market, and my goals well enough to give me something genuinely useful.”
AI then asks questions like:
“What type of recruitment service is this, and how does it differ from what you currently offer?”
“Who are your ideal clients and what size companies are we talking about?”
“What do your competitors charge for similar services?”
“What are your costs for delivering this service?”
“What profit margin do you need this to achieve?”
“Have you tested any pricing with potential customers? What was their reaction?”
“What’s your positioning—premium, competitive, or value-focused?”
With voice responses, the owner naturally provides rich context: “So the new service is a retained search for senior operations roles, we’ve been doing contingency recruitment for years but this is different because... our main competitor charges around 25% but they’re more established so... we did mention ballpark figures to two prospects who seemed okay with...”
After this interview, AI has genuine context. The pricing strategy it produces is specific to this business, this market, this competitive landscape, this positioning.
Same business owner. Same AI. Radically better output.
Where This Works Best
This interview approach is particularly valuable for strategy and planning tasks.
Business strategy decisions where context matters more than generic frameworks. Marketing plans that need to account for your specific situation. Pricing strategies that depend on your costs, market, and positioning. Hiring decisions where culture and team dynamics are relevant.
Basically, any task where “it depends” is the honest answer to most questions. The interview extracts what it depends on in your case.
It’s less necessary for straightforward tasks. “Write a professional email declining this meeting” doesn’t need a deep interview. The context is obvious.
But for the meaty stuff—the decisions and plans that actually drive your business—the interview approach consistently outperforms traditional prompting.
Three Starter Prompts
Here are copy-paste prompts for common strategy tasks:
For Marketing Strategy:
“I need to develop a marketing strategy for the next quarter. Before you create anything, interview me about my business, my audience, my current efforts, and my goals. Ask questions until you have enough context to give me something genuinely tailored to my situation.”
For Pricing Decisions:
“I’m trying to figure out the right pricing for [product/service]. Rather than giving me generic pricing models, interview me about my costs, my market, my competitors, and my positioning. Keep asking questions until you understand enough to give me specific recommendations.”
For Business Planning:
“I need to create a plan for [specific initiative]. Before you draft anything, ask me questions to understand my business context, constraints, resources, and what success looks like. I want your output to be specifically tailored to my situation, not generic advice.”
Why This Bypasses Prompt Engineering
Prompt engineering is essentially the skill of anticipating what context AI needs and providing it upfront.
The interview approach sidesteps this entirely. You don’t need to know what’s relevant. You don’t need to structure information correctly. You don’t need to use specific frameworks or magic words.
You just need to answer honestly when asked.
The AI figures out what questions to ask. You provide the answers. The AI uses those answers to give you something useful.
This is accessible to anyone. No special skills required. No courses on “how to write better prompts.” Just a willingness to have a conversation.
The Practical Bit
Try this today.
Pick something you need AI help with. A decision you’re wrestling with. A plan you need to create. A strategy you’re developing.
Instead of crafting a detailed prompt, use this simple format:
“I need [what you’re trying to achieve]. Before you create anything, interview me until you understand my situation well enough to give me something genuinely useful.”
If you have a voice-to-text tool like Wispr Flow, use it for your responses. Talk through your answers rather than typing them. Let yourself elaborate.
Notice the difference in output quality.
Once you experience it, you won’t go back to the old way of cramming context into long prompts and hoping AI figures it out.
The Bigger Picture
The goal of using AI well isn’t to become a prompt engineer. It’s to get useful output with minimal friction.
The interview approach achieves this by putting AI in the driver’s seat for context gathering. It knows what questions to ask. You know the answers. Together, you produce something that actually fits your situation.
This is the 10/80/10 framework in action. 10% of your effort is explaining what you need and answering questions. 80% is AI doing the work with proper context. 10% is refining the output.
Here’s what works: letting AI ask you questions rather than guessing what information to provide upfront. What doesn’t: spending ages crafting prompts and still getting generic outputs.