Crafting Effective Prompts – Structure, Context, and Constraints Log in and enroll to track lesson completion. By the end of this lesson, students should be able to: The cost of wrong-direction output. When Claude misunderstands your intent and produces a long, detailed, well-written response to the wrong question – that is still wasted time. You have to re-read it, figure out what went wrong, and restart. For short tasks, this is minor. For long documents, complex analysis, or multi-step plans, a wrong-direction output is a significant setback. The solution is not to write longer prompts every time. The solution is to use two powerful techniques that catch misalignments before Claude commits to an answer. Technique 1: Ask Claude to state its assumptions. Add this to any complex prompt: "Before you answer, list the key assumptions you are making about my request." Claude will surface the interpretations it chose when your prompt was ambiguous – and you can correct them before it proceeds. This is especially valuable when: Example addition to any prompt: "Before answering, briefly state what you are assuming about [my role / the audience / the scope / the technical context]. I will confirm or correct before you proceed." Technique 2: Ask Claude to request clarification first. For complex, high-stakes, or genuinely ambiguous requests, you can explicitly tell Claude: "Before you begin, ask me any clarifying questions that would help you give a better answer. Wait for my response before proceeding." This inverts the usual pattern. Instead of Claude guessing and you correcting, Claude asks and you inform. The result is almost always a more targeted first output. Use this technique when: Technique 3: Review Claude's reasoning, not just its conclusion. Ask Claude to "think through this step by step" or "show your reasoning." This lets you spot where Claude went wrong before accepting the output. A wrong conclusion from correct reasoning tells you something different from a wrong conclusion from flawed reasoning – and you cannot tell the difference if you only see the output. When not to use these techniques. For simple, low-stakes tasks – "summarize this paragraph," "rephrase this sentence," "suggest five names for this product" – assumption-surfacing and clarification requests add friction without benefit. Use them for complex, high-stakes, or ambiguous prompts where a wrong direction is costly. A consultant needs Claude to help write a competitive positioning memo. Weak approach: Write a competitive positioning memo for my company. Claude has no idea what the company does, who the competitors are, what the positioning angle should be, or who will read the memo. Better approach with assumption-surfacing: I need help writing a competitive positioning memo. Before you begin, list the assumptions you are making about the company type, the competitive landscape, the audience, and the intended use of this document. I will confirm or correct before you proceed. Better approach with clarification request: I need help writing a competitive positioning memo. Before you start writing, ask me the questions that would most improve your output. Wait for my answers before drafting anything. Both approaches prevent Claude from producing a generic template that requires complete reconstruction. You can use assumption-surfacing as a diagnostic tool when Claude keeps producing outputs that miss your intent. If you have iterated three times and Claude keeps going in the wrong direction, ask: "Before we continue, what do you currently understand my goal to be? What assumptions are you working from?" Claude's answer often reveals a fundamental misalignment in how it has been interpreting the task – which you can then correct directly. When Claude asks clarifying questions before answering, the questions themselves sometimes reveal gaps in the original prompt that have privacy or safety implications. If Claude asks "what is the legal jurisdiction for this question?" it is signaling that your request has legal dimensions that vary by location – and that you should be thoughtful about treating any answer as authoritative without consulting an appropriate professional. Write a prompt for a complex task you are working on. First, submit it as-is and note the output. Then add "Before you answer, list the key assumptions you are making about my request." Compare the two outputs. How did the surfaced assumptions change your understanding of what Claude was inferring? You should be able to add assumption-surfacing or clarification-request language to any complex prompt, and explain why this technique reduces wrong-direction output. Log in and enroll to take this lesson quiz.
Lesson 4: Asking Claude to Surface Assumptions and Request Clarification
Lesson Objectives
Lesson Content
Practical Example
Lesser-Known Tip
Safety Notes
Practice Task
Completion Check
Lesson Quiz