Verified against ChatGPT · 2026-08-13
Build an interview podcast episode outline that gets a guest to the good stories without a rigid Q&A script
Produces a flexible outline organized around the specific stories and tensions worth surfacing from a guest, with built-in follow-up branches, instead of a linear list of questions to read off.
The prompt
Ready to copy — highlighted parts are example details you can swap.
You are building an outline for an interview-format podcast episode. This is not a script of questions to read verbatim — it's a map of the territory worth covering, with follow-up branches, so the host can actually listen and react instead of just running down a list. GUEST AND WHY THEY'RE INTERESTING A former commercial airline pilot who now investigates near-miss incidents; interesting because he has stories his current employer won't let him tell publicly, only ones from years ago. EPISODE ANGLE How small procedural shortcuts, not dramatic failures, cause most near-misses. WHAT THE AUDIENCE ALREADY KNOWS ABOUT THIS GUEST OR TOPIC Regular listeners already know what a near-miss report is from a previous episode; no need to re-explain the basic concept. THE ONE STORY OR TENSION WORTH GETTING TO A specific incident he's mentioned in passing on Twitter but never told in full, involving a shortcut he took under time pressure. PHASE 1 - OPENING TERRITORY (first 5-8 minutes) Draft 2-3 opening questions designed to get the guest talking naturally and establish context the audience needs, not generic icebreakers. For each, note one likely follow-up direction depending on how they answer. PHASE 2 - THE MIDDLE, BUILT AROUND THE KEY STORY Map out the sequence of questions or prompts most likely to lead naturally to the key story or tension named above, without asking for it directly and bluntly if that would make the guest perform rather than actually tell it — build toward it the way a real conversation would, through an adjacent, easier question first. Include at least one planned follow-up branch for if the guest gives a short or guarded answer to the lead-in. PHASE 3 - THE HARD OR SPECIFIC QUESTION Write the direct version of the question that gets at the key story or tension, positioned for where in the conversation it will land best, plus a fallback rephrasing if the first version gets deflected. PHASE 4 - CLOSE One or two questions that give the guest a natural way to wrap (advice, what's next, where to find them), avoiding a flat 'anything else you want to add.' WHAT NOT TO DO Do not write this as a linear numbered question list with no branching — a real conversation doesn't go in a straight line, and an outline that pretends it will forces the host to either abandon the outline entirely or awkwardly steer the guest back to the script. Do not front-load the hardest question before there's any rapport. OUTPUT FORMAT The four phases as labeled sections, each question followed by its follow-up branch(es) indented beneath it, so the host can see the whole decision tree at a glance rather than just a flat list.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
A guest who feels like they're being run through a checklist tends to give shorter, more rehearsed answers than one who feels like they're in an actual conversation, because a visible linear script signals to the guest exactly where the conversation is headed and invites a prepared, safer answer rather than the more revealing one that comes from feeling genuinely listened to — which is precisely the difference between an outline with follow-up branches and the flat numbered question list GPT-5.1 defaults to when simply asked for 'interview questions.' Structuring the middle phase to approach the key story through an easier adjacent question first, rather than asking for it directly and early, mirrors how skilled interviewers actually work: a guarded story told on demand reads as a rehearsed anecdote, while the same story reached through a natural conversational on-ramp tends to come out with more specific, unplanned detail, because the guest is answering the question actually in front of them rather than performing for an anticipated one. Requiring a fallback rephrasing for the direct question addresses a concrete, common failure in unscripted interviews — a guest deflecting or giving a short non-answer to the first phrasing — that a single-question outline leaves the host with no prepared way to handle, forcing an improvised follow-up in the moment that's often weaker than one planned in advance. The explicit instruction against a flat linear list matters because a numbered list is the model's statistical default for 'outline,' and using that shape here would produce something that looks organized but actively works against the goal, since a host trying to follow it verbatim ends up steering the guest back onto rails instead of following what's actually said.
What you get back
Phase 2 (excerpt): Q - 'What's the most common cause of a near-miss report, in your experience — not the dramatic version people imagine?' Follow-up branch if he gives a general/procedural answer: 'Was there a specific shift where you saw that play out?' Follow-up branch if he mentions time pressure specifically: 'You've mentioned before that time pressure got you into a tight spot once yourself — is that the kind of thing you mean?'
Verified against
ChatGPT GPT-5.1 · 2026-08-13
Changelog
- 2026-08-13 — Initial publish, verified against ChatGPT GPT-5.1.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
