Every host has done the version where you wing it. Guest joins, you exchange pleasantries, you glance at three questions you scribbled that morning, and forty minutes later you realize the best thing they said got no follow-up because you were already looking at question four.

Preparation is what stops that. Not the kind that turns into a rigid script you read off a page, but a system you can run in an hour or two and repeat for every guest. This guide is for hosts who already record interviews and want the conversation to go somewhere on purpose. It walks through researching a guest, building questions you can actually adapt, mapping the episode, and staying present enough to chase the good stuff when it shows up.

Start with the research, but bound it

Guest research is where most advice starts, and it's right to. Read what they've written, skim their site, and — this is the part people skip — listen to one or two other interviews they've done. That last step does two jobs. It tells you the stories they always tell (so you can skip them or dig past them), and it flags the questions everyone asks so you can ask something else.

Set a limit, though. An hour of research for a one-hour episode is plenty. It's easy to fall into a rabbit hole reading a guest's entire back catalog and show up with forty pages of notes you'll never use. What you actually need is small: three or four things they care about, one thing they've said that you disagree with or want to push on, and a couple of specifics — a project, a number, a claim — you can reference by name so they know you did the work.

Write that into a one-page brief. Name, what they're known for, the angle for this episode, and the handful of moments you want to reach. If you can't fit it on one page, you haven't decided what the episode is about yet.

Build a question ladder, not a script

Here's the shift that changes interviews: stop writing a list of questions and start writing a ladder.

A list is flat. Question one, two, three, and you march through them regardless of what the guest says. A ladder is ordered by depth. The bottom rungs are easy, factual, warm-up — "how did you get into this?" The top rungs are the ones you actually care about, the ones that need trust and momentum before a guest will answer them honestly.

The reason this matters: you rarely get to the top of the ladder by asking directly. You get there because a warm-up answer cracked a door open and you walked through it. So write the ladder as themes with a lead question each, plus two or three follow-ups you'd love to reach — and hold them loosely. Six to eight themes is usually enough for an hour. More than that and you're back to marching.

One practical rule: for every theme, know the one question you'd regret not asking. Star it. If the conversation runs long and you have to cut, you cut around the stars, not through them.

Decide what you're sending the guest ahead of time

Sending questions in advance is one of those debates hosts never settle. Send everything and the answers can come out rehearsed and flat. Send nothing and a guest can freeze, or spend the first ten minutes finding their footing on tape.

The middle works best. Send the themes and one or two specific questions you want them to prepare for — the "bring your three favorite tools" type, where a considered answer beats a spontaneous one. Keep the follow-ups and the sharper questions to yourself. Your guest arrives oriented, not scripted, and the interesting turns still happen in real time.

Tell them the practical stuff too: how long you'll record, whether it's video, that you edit out stumbles so they should feel free to restart a sentence. A guest who knows they can flub a line relaxes, and relaxed guests are better guests. The full list of what to send a podcast guest before recording — and the point in the week each piece should land — is worth setting up once and reusing. That confirmation note also does quiet double duty: it's the single best predictor of whether your guest shows up on time — or at all.

Map the run-of-show

The run-of-show is a rough clock for the episode. Cold open or straight into it? A standard intro you read every time? Where does the meat sit, and roughly how many minutes does each theme get?

You will not follow it exactly. That's fine — its job isn't to be obeyed, it's to keep you from looking up at minute fifty realizing you spent thirty minutes on the warm-up and have eight left for the reason you booked this guest. A loose clock in your peripheral vision fixes that. Budget time per theme, protect the starred questions, and let everything else flex. If you want the actual document, with minute markers and a rule for what gets cut when you run long, here is how to write a podcast run of show. And when a recording goes sideways despite the plan, that same run-of-show is what makes rescuing a bad interview a navigation problem instead of an emergency.

Where an AI co-pilot fits into prep

A lot of the prep above is mechanical, and that's exactly the part worth handing off.

Liyam AI takes a topic and a few links about your guest and turns them into a structured plan — an episode outline with talking points and suggested questions tailored to that guest. It's a starting draft, not the final word. You still edit it like an owner: cut the questions that don't sound like you, reorder so the conversation builds, and delete anything you already know the answer to. But it collapses the blank-page part of prep, which is where the hours quietly go.

The bigger shift is during the recording. A live co-pilot listens alongside you: when a guest cites a statistic, it can check the claim in seconds; when a thread opens up, it offers follow-up questions so you can chase it without losing your place; and it nudges you toward the next theme when one runs long. That's what lets you prep a ladder instead of a script in the first place. You don't have to hold every branch of the conversation in your head, so you can actually listen — which is the whole job.

Afterward, the slow parts move fast: an accurate transcript within minutes, a drafted summary and show notes to edit, and the standout moments flagged for the clips you'll cut. Prep, in other words, doesn't end when you stop recording. Good prep sets up a fast finish.

The last hour before you record

Prep degrades if you cram it all in earlier and then walk in cold. Save a short block right before:

  • Reread your one-page brief and the starred questions. Not the whole ladder — just the spine.
  • Test the recording setup. Mic, levels, backup recording on. The best interview is worthless if the audio isn't.
  • Kill the distractions. Phone on airplane mode, notifications off, other tabs closed. A single Slack ping mid-answer and you've lost the thread.
  • Have your first two questions ready cold. The opening is the one moment you can't improvise well, because you and the guest are both still warming up.
  • Pick your closing question while you are at it. The opening and the close are the two moments worth deciding in advance, and ending the interview is its own skill: the last ten minutes are where good episodes get wasted.

Then close the notes and talk to the person.

FAQ

How long should it take to prepare for a podcast interview?

For a weekly show with a familiar format, roughly one to two hours per episode once you have a repeatable system — most of it guest research and building the question ladder. Your first few will take longer. The system is what brings it down.

Should I send interview questions to my guest beforehand?

Send the themes and one or two questions you want them to prepare thoughtful answers to. Keep your follow-ups and sharper questions for the live conversation. That gives the guest confidence without making the whole thing sound rehearsed.

How many questions should I prepare?

Think in themes, not a long question list. Six to eight themes for a one-hour episode, each with a lead question and a couple of follow-ups, is plenty. Star the one question per theme you'd regret skipping.

What if the conversation goes off-script?

Let it, within reason. The off-script moments are usually the best ones. A loose run-of-show and your starred questions are the safety net — glance at the clock, protect the must-asks, and follow the tangent as long as it's earning its place.

Bottom line

Preparing for a podcast interview isn't about controlling the conversation. It's about earning the freedom to let go of it. Do the research, build a ladder you can climb in any order, map a rough clock, and set up your tools so the busywork isn't stealing your attention from the guest.

Want the prep and the cleanup handled so you can focus on the talking? See how Liyam AI plans your episodes and rides shotgun while you record.