Think in layers when using AI

Working with transcripts can feel overwhelming. Client calls. Workshop recordings. Interview transcripts. Pages and pages of raw material—good ideas buried under tangents, half-finished thoughts, and off-the-cuff remarks. The problem usually isn't lack of content. It's too much of it! In this episode, I walk you through a simple, repeatable workflow I use to turn messy transcripts or rough notes into a clear, usable outline—without losing the nuance that actually matters. If you've ever dropped a transcript into AI, asked it to "summarize this," and felt underwhelmed by the result… this episode will show you a much better approach. What You'll Learn Why asking AI to "summarize" is usually the wrong first move How to give AI better signal by starting with context, not content A practical, copy-and-paste prompt for structuring messy transcripts How to preserve nuance, tension, and unresolved thinking Where AI's role ends... and where your judgment matters most Key Ideas Covered in This Episode 1. The Real Risk of AI Summaries AI summaries are often: Clean Organized And emotionally flat When you ask AI to summarize too early, it tends to: Smooth over tension Resolve ambiguity prematurely Erase the very moments that make the thinking interesting But those messy moments are often the most valuable parts of a transcript! 2. Start With Context Before Content Before pasting anything into AI, clarify: What this material is (a client call, interview, workshop, etc.) What you're trying to create (article outline, memo, talk, case study) Who it's for What matters more here: clarity, persuasion, or depth This framing alone dramatically improves the output. 3. Don't Hide Your Own Thinking If you were part of the conversation—or listening closely—you already have insight. You noticed: Patterns Tensions Strong opinions What felt important (even if you're not sure why yet) Dump that thinking into the chat. You can literally say: "Here are my rough and random thoughts so far. None of this is locked in." That gives the model far better signal than a raw transcript on its own. 4. Ask for Structure—Not Writing Before asking AI to write anything, ask it to: Identify themes and recurring ideas Group related concepts into buckets Flag contradictions or unresolved thinking Preserve nuance instead of smoothing it out You're looking for a skeleton here. That's it. 5. A Simple Prompt You Can Use Here's the exact type of instruction I recommend at this stage: Take the role of a skilled research assistant helping me make sense of raw thinking without oversimplifying it. Study the transcript I've attached, along with my rough notes and early thoughts. Nothing here is finalized. I need you to: · Identify the main themes, tensions, and recurring ideas · Group them into a clear outline · Flag nuance, contradictions, or unresolved thinking · Do not write prose, conclusions, or clean summaries This keeps the AI in the right lane. 6. Where Your Role Becomes Clear Once you have structure: You decide what stays You decide what moves You decide what gets cut or combined AI gives you a map. Now it's up to you to choose the route. At this point, writing becomes easier. Not because AI wrote it for you, but because the thinking is no longer chaotic. The Big Takeaway Think in layers. So instead of asking AI to finish the job in one move, use it to: Identify patterns Clarify structure Reduce cognitive load When used this way, AI amplifies your judgement. And that's the goal: to let smart tools handle the grunt work so you can focus on framing, meaning, and persuasion. Listener Reflection Here's the question I'll leave you with: What part of your current workflow would benefit most from letting AI point out patterns while you keep the final call? If you found this episode useful, make sure you're subscribed so you don't miss future conversations on using tools thoughtfully, without giving up your edge as a creative professional.