The Gap Between Your Intent and the AI Output
When you ask an AI to write a report or plan a trip, the results often feel generic or slightly off. Most people try to fix this by writing longer, more complicated instructions. They spend minutes trying to think of every possible detail the AI might need, yet they still forget the most important context. This creates a cycle of frustration where you spend more time fixing the AI's mistakes than if you had done the work yourself.
The real problem is not your writing style; it is a lack of shared context. Context is the collection of background details, personal preferences, and specific goals that live in your head but not in the chat box. Instead of guessing what the AI needs to know, you can use a simple shift in strategy: make the AI do the work of finding that information for you.
The Interview Prompt Technique
The most effective way to start any project is to tell the AI to interview you. Instead of saying 'write a three-day itinerary for London,' say 'I want to plan a trip to London. Interview me one question at a time to learn about my budget, interests, and schedule before you suggest anything.' This forces the AI to identify what information is missing. By answering one question at a time, you provide high-quality data without getting overwhelmed by a massive list of requirements.
This recursive questioning ensures that the AI maps out your project goals before it starts generating any final content. It prevents the assistant from making assumptions about what you want. If you are planning a dinner party, the AI might ask about dietary restrictions you forgot to mention. If you are drafting a business email, it might ask about the history of your relationship with the recipient. These small details are what make the final result feel accurate and useful.
Building a Project Map Before the Final Draft
Once the interview is finished, do not rush into the final product. Ask the AI to create a Project Map based on your answers. This is a summary of everything it has learned and a step by step plan for the final task. In professional settings, this is often called an implementation spec, but for everyday use, think of it as a checkpoint. It allows you to see if the AI truly understands your needs before it spends time writing hundreds of words.
Reviewing this map gives you a chance to correct misunderstandings early. If the AI suggests a formal tone for a casual event, you can fix that instruction in the map. This layer of verification saves time because you are editing a short outline rather than a long, completed document. Only when the Project Map looks perfect should you give the AI the signal to generate the final version of your project.
Adding Your Personal Taste to the Final Result
The greatest strength of the interview method is that it captures your personal taste. AI is excellent at structure and grammar, but it lacks your unique perspective. By answering specific questions during the discovery phase, you are injecting your own decision making logic into the conversation. This ensures the output reflects your standards and style rather than a generic average of the internet.
You can even tell the AI to remember these preferences as a new skill for future chats. For example, if you realize you always prefer short bullet points over long paragraphs, tell the AI: 'Based on this conversation, remember that I prefer concise updates for all future tasks.' This turns a single successful interaction into a repeatable habit, making the AI smarter and more helpful every time you use it.
Key takeaways
- Use the phrase 'interview me one question at a time' to let the AI find the context it needs.
- Always ask for a Project Map or summary before the AI begins the final task to catch errors early.
- Provide specific preferences about tone and style during the interview to ensure the results match your personal taste.
- Focus on one question at a time to prevent the AI from making generic assumptions about your goals.
- Tell the AI to save your preferences as a 'skill' so you do not have to repeat yourself in future conversations.