Large language models (LLMs) excel at handling complex tasks, but they often require extensive context—detailed descriptions, implementation guidelines, and external data sources. Traditionally, humans write this context manually. But there's a smarter way: let the LLM interview you. By prompting the model to ask targeted questions, you can efficiently transfer your knowledge into a structured document. This technique, known as an interrogatory LLM, not only streamlines context creation but also helps review existing documents and assists people who struggle with writing. Below are five powerful strategies to put this method into practice.
1. Build Context from Scratch with an LLM Interview
When designing a new feature or solving a complex problem, you often need to provide an LLM with a rich set of background information—user interface requirements, technical constraints, integration details, and more. Instead of spending hours crafting a multipage document yourself, you can instruct the LLM to interrogate you. It will ask a series of questions to gather all necessary details. You answer verbally or in short written responses, and the LLM compiles everything into a coherent context report. This report can then be used in a separate session (perhaps with a different model) to generate the final output. The result is a faster, more interactive way to feed an LLM the context it needs, ensuring nothing is overlooked.
2. Adopt the One-Question-at-a-Time Approach
Harper Reed, who popularized this technique, emphasizes a crucial rule: the LLM should ask only one question at a time. This prevents information overload and keeps the conversation focused. When you try it, you may need to repeatedly remind the model to stick to this rule—LLMs naturally tend to fire a barrage of queries. By enforcing a single-question cadence, you give yourself time to think deeply about each answer, leading to higher-quality context. This step-by-step interrogation also mirrors human coaching, making the process feel more natural and less like a data dump. It’s a small adjustment that significantly improves the reliability and completeness of the generated context.
3. Review Documents Through Conversational Audits
Interpretation of existing documents—like software specifications—can be tedious and error-prone when done alone. Instead of asking a human expert to read and critique a dense paper, hand the document to an LLM and instruct it to interview the expert. The LLM will ask questions about the document’s accuracy, completeness, and clarity. This conversational approach often uncovers issues that a silent read might miss, especially if the document is poorly written. Experts find it easier to answer questions in a dialogue rather than dissecting a text. The LLM then produces a summary of findings, making the review process more engaging and effective. It’s a win-win: the expert remains engaged, and the review becomes structured.
4. Combine Interviewing for Both Creation and Review
The true power emerges when you chain these techniques. Use one interrogatory LLM to build a context document by interviewing the initial knowledge holder. Then, deploy a second interrogatory LLM to review that document with another expert—or even with the same expert for validation. This creates a two-step loop: creation followed by verification. You can repeat the process iteratively until the document meets your standards. This approach leverages the model’s ability to ask consistent, thorough questions, while human experts provide nuanced answers. It’s especially useful in high-stakes scenarios like regulatory compliance or system architecture, where accuracy is paramount and human error must be minimized.
5. Help Non-Writers Share Their Knowledge
Not everyone enjoys writing. For many people, getting their thoughts into a clear document is a painful, slow process. This can block critical knowledge from being captured. An interrogatory LLM offers a lifeline: instead of writing a report, a person can simply answer the model’s questions in a dialogue. The LLM then converts those answers into a well-structured document. Yes, the output may carry a hint of AI-generated prose, but that is far better than having no documentation at all—or a rushed, incomplete one. This technique democratizes knowledge sharing, enabling subject matter experts who dread writing to contribute their insights effectively. It transforms knowledge extraction from a solitary chore into a collaborative conversation.
An interrogatory LLM shifts the dynamic from human-to-machine typing to human-to-machine conversation. Whether you’re building context from scratch, reviewing documents, or capturing tacit knowledge, letting the LLM lead the questioning saves time and improves quality. Start with one strategy, and you’ll soon discover how natural and effective it is to have a model interview you. The future of human-AI collaboration is not just about giving commands—it’s about having a dialogue.