AI Skill Report Card

Creating Carousel Content

A-86·Nov 21, 2026·Source: Web
14 / 15

Given raw media (photos/videos from a phone or camera), immediately:

  1. Identify the 3-5 strongest visual assets (best lighting, composition, emotion, or "scroll-stopping" quality)
  2. Determine the narrative arc across slides (problem→solution, before→after, myth→truth, process reveal, etc.)
  3. Draft the hook (slide 1) — this determines 80% of performance
  4. Sequence remaining slides to maximize swipe-through rate
  5. Write caption + CTA
  6. Flag one "0.1% move" — something top creators would do that's non-obvious

No preamble. Jump straight into the carousel breakdown.

Recommendation▾
Add a brief handling note for when assets are genuinely weak/insufficient (e.g., only 1-2 usable photos) rather than assuming always 3-5+ strong assets.
14 / 15

Progress:

  • Step 1: Audit the raw assets (quality, variety, usable angles)
  • Step 2: Identify the brand's current content gap or opportunity
  • Step 3: Pick a carousel format/archetype
  • Step 4: Draft slide-by-slide breakdown
  • Step 5: Write hook + caption + CTA
  • Step 6: Challenge the creator with a sharper alternative
  • Step 7: Note what top 0.1% accounts do differently here

Step 1: Audit the Assets

Categorize what's provided:

  • Photos: static, high-res, good for text-overlay slides, data/quote slides, product detail
  • Videos: better for slide 1 (motion stops scroll even in carousels—use a looping cover frame) or as a "swipe to watch" final slide
  • Note what's missing (e.g., no face/personality shot, no clear CTA visual)

Step 2: Identify the Gap

Ask: what has this brand NOT posted recently? Options:

  • Personality/behind-the-scenes (builds trust)
  • Educational/listicle (builds authority)
  • Social proof/results (builds credibility)
  • Opinion/hot take (builds engagement + reach)

Pick based on what's underrepresented in their recent content, not what's easiest.

Step 3: Pick a Carousel Archetype

Default options (pick ONE, don't hedge):

  • The Listicle: "5 things nobody tells you about X" — each slide = 1 point
  • The Myth-Bust: Slide 1 = common belief, Slides 2-4 = why it's wrong, Slide 5 = truth + CTA
  • The Story Arc: Slide 1 = relatable problem, middle = journey, last = resolution/product tie-in
  • The Single Idea, Expanded: One strong insight, each slide zooms deeper
  • The Visual Diary: For personality-heavy brands — looks unplanned but is tightly curated

Step 4: Slide-by-Slide Breakdown

For each slide, specify:

  • Which asset (photo/video frame) goes here
  • Text overlay (exact copy, not placeholder)
  • Why this slide earns the swipe to the next

Step 5: Hook + Caption + CTA

  • Hook (slide 1 text): Must create a curiosity gap or call out the viewer directly. Avoid generic ("Check this out!"). Aim for specificity + stakes.
  • Caption: First line = hook continuation (most captions get cut off). Body = context/value. End = soft CTA (comment, save, share) — not just "link in bio."
  • CTA: Match to goal — saves for educational, comments for opinion, shares for relatable/funny.

Step 6: Challenge the Creator

Always push back on the obvious choice. Examples of challenges to raise:

  • "This hook is safe. What's the version that risks a few people disagreeing?"
  • "You're leading with the best photo — top accounts often bury it at slide 3 to force a swipe."
  • "This caption explains the carousel. The carousel should do that; the caption should add a layer (opinion, story, or counterintuitive detail) they can't get from slides alone."

Step 7: The 0.1% Insight

Always include ONE non-obvious tactic top creators use, e.g.:

  • They design slide 1 and the LAST slide as a pair — the last slide often loops thematically back to the first to reward full swipe-throughs.
  • They sacrifice "nice" photos for photos with visible imperfection/authenticity — polish reads as ad, texture reads as real.
  • They write captions as if replying to a comment, not announcing a post — this unconsciously invites comments.
  • They A/B the hook text only, keeping the same visuals, and track saves not likes.
Recommendation▾
Include a third example with a different brand vertical (e.g., B2B/SaaS) to show the framework generalizes beyond lifestyle/fitness brands.
18 / 20

Example 1: Input: 6 iPhone photos from a café brand's new matcha launch — product shots, a barista pouring, a candid customer laughing, an empty cozy interior shot.

Output:

  • Archetype: Story Arc
  • Slide 1: Candid customer laughing (text: "POV: you ordered the 'basic' matcha and regretted doubting us") — hook via relatable sarcasm, not product-first
  • Slide 2: Barista pour (text: "it's whisked, not shaken — here's why that matters")
  • Slide 3: Close-up product shot (text: "3 ingredients. no syrup needed.")
  • Slide 4: Empty interior (text: "also... best seat in town is still free at 3pm")
  • Slide 5: Product shot again, different angle (text: "tag someone who needs this today")
  • Caption: "We didn't plan for the matcha to outsell the espresso. Here's what's actually in it 👇"
  • 0.1% insight: Slide 4 (interior) seems off-topic for a product launch — but it's the slide most likely to get saved by locals planning a visit, which signals "save" to the algorithm more than any product shot would.
  • Challenge: "Your instinct was to open with the product shot. That's what every café does. Opening on the laughing customer forces curiosity — what happened? — before they even know it's about matcha."

Example 2: Input: A fitness coach sends 10 gym video clips, mostly form demonstrations.

Output:

  • Archetype: Myth-Bust
  • Slide 1 (video, looping clip of a common mistake in slow-mo): "You've been doing this wrong for years"
  • Slides 2-3: Freeze frames with text overlay pointing out the error
  • Slide 4: Correct form video clip
  • Slide 5: Text-only slide: "Save this before your next leg day"
  • 0.1% insight: Top fitness creators often show the WRONG way first in full confidence (no "don't do this" label) for 1-2 seconds — it bait-checks viewers who'll comment correcting them, which the algorithm rewards as engagement, before slide 2 reveals it was intentional.
  • Challenge: "You wanted to lead with the correct form to look authoritative. But correct-form content is everywhere. Leading with the mistake creates tension — that's what gets saved and shared, not competence alone."
Recommendation▾
Consider trimming some repetition between 'Best Practices' and 'Common Pitfalls' sections, as several points overlap (e.g., leading with polished asset vs. best asset).
  • Always give an exact hook line, never "write something punchy here."
  • Keep text overlays under ~8 words per slide — carousels are skimmed, not read.
  • Odd slide counts (5, 7) tend to feel more intentional than even counts.
  • Match content archetype to brand's current gap, not creator convenience.
  • Always deliver ONE clear direction, not three options — creators want decisiveness, not a menu.
  • Tie the "0.1%" insight to something specific and tactical, never generic ("be authentic").
  • Don't lead with the most polished/"best" asset by default — it's often better as a mid-roll payoff.
  • Don't write captions that just restate the slides — they should add a missing layer.
  • Don't suggest more than one carousel concept unless asked — commit to the strongest one.
  • Don't give vague challenges ("push yourself creatively") — always name the specific safe choice and the sharper alternative.
  • Don't ignore video assets by treating them as unusable in carousels — a short looping clip as slide 1 or the final slide outperforms static images.
  • Don't default to "link in bio" as the only CTA — vary by content goal (saves/comments/shares).
0
Grade A-AI Skill Framework
Scorecard
Criteria Breakdown
Quick Start
14/15
Workflow
14/15
Examples
18/20
Completeness
17/20
Format
14/15
Conciseness
13/15