Why do AI carousels all look the same?

Because most tools generate from a prompt about a topic rather than extracting from what the creator actually said. The source material is the same generic internet knowledge base, not the creator's unique teaching.

How many slides should a LinkedIn carousel have?

Between 7 and 12 slides. Under 7 and there is not enough substance to justify the swipe. Over 12 and you are padding. The sweet spot is 8–10 slides with one clear idea per slide.

What makes a carousel hook work?

A hook that names a specific problem, outcome, or number the audience recognizes immediately. Vague benefit hooks ("improve your content") lose to specific tension hooks ("why your carousels get saved but never shared").

The Anti-Generic Guide to AI Carousels

September 30, 2026 By Aran P Ploshansky

AI carousels have a sameness problem. How to build LinkedIn carousels from your podcast that sound like you — hook, structure, and design rules included.

Scroll LinkedIn for five minutes and you can spot them instantly: the AI carousels. Same hook formulas. Same "5 lessons" structure. Same vaguely motivational close. The tooling that made carousels easy to produce also made them easy to ignore.

The problem isn't AI. The problem is *where the content comes from.* Generic carousels are generated from nothing — a topic and a prompt. Great carousels are extracted from something — a conversation where you actually said something specific.

## The source-material rule

Never build a carousel from a topic. Build it from a moment.

When you [repurpose from a podcast episode or video](/blog/turn-one-podcast-episode-into-30-days-of-content), you're not asking a model to have your ideas. You're asking it to *find* your ideas — the three minutes where you explained something better than you ever have in writing. That specificity is the entire difference between a carousel that gets saved and one that gets scrolled.

## Anatomy of a voice-preserving carousel

**Slide 1 — The hook is a claim you actually made.** Not "5 Content Mistakes to Avoid." Instead, the sharpest sentence from the episode, close to verbatim: "Your best content is trapped in your worst format." If it made your co-host pause, it'll make the feed pause.

**Slides 2–3 — The felt problem.** Before the framework, the frustration. Describe the reader's situation in the language you used on the show. This is where flattened AI writing dies and real phrasing survives.

**Slides 4–7 — The insight, in your structure.** Resist the urge to reformat everything into tidy numbered lists. If you explained it as a story on the episode, keep it a story. Structure should follow the source, not the template.

**Slide 8 — One CTA.** Not three. One. Ours is always a variation of the same line: turn your next episode into a content batch.

## Design rules that protect the voice

- **Verbatim beats paraphrase.** If a slide quotes you, quote you exactly — including the odd phrasing. Odd phrasing is a fingerprint. - **Your brand system, every asset.** Fonts, colors, [layouts and aspect ratios](https://hyrstudio.com/features/styles-aspect-ratios). A carousel that could belong to any account belongs to no one. - **Cut the filler slides.** "Let's dive in" is a slide someone swipes past. Every slide earns its swipe or gets cut.

## The batch advantage

One episode usually yields 3–5 carousel-worthy moments — different insights for different audience segments. Generating them [as a batch](/blog/how-we-use-hyr-studio-to-market-hyr-studio) (rather than one artisanal carousel per week) means you can A/B hooks, spread themes across the month, and keep the strongest opener for your biggest publishing day.

That's the actual promise of AI in this workflow: not writing for you, but mining you — at a volume you'd never manage by hand.

**Turn your next episode into a content batch →** [hyrstudio.com](https://hyrstudio.com)