I read about ScrollEd on TechCrunch this morning and had one of those moments where you nod along thinking "yes, obviously, why didn't more people do this sooner."
The pitch is simple. Take static, boring textbook content — the stuff students already own but never actually read properly — and chop it into short, scrollable video segments with quizzes baked in. Basically TikTok's format applied to GCSE biology instead of dance routines. Founded by students, built for students, and apparently sticky enough that kids are choosing to revise with it rather than avoid it.
I've been building tools that do version of this at scale for years — Article2Video turns written content into narrated video, MarketMasher chews through source material and spits out multiple content formats from one input. So when I see a scrappy student team doing the same fundamental job by hand, at micro-scale, my instinct isn't to dismiss it. It's to ask what they've got right that a lot of well-funded automation plays get wrong.
The Format Is the Product
The thing ScrollEd understands, which a lot of "content repurposing" startups miss, is that the format isn't decoration on top of the content — the format is the product. A textbook chapter chopped into a video with the same words in the same order is just a boring video. What makes this work is the quiz-lacing: forcing a decision, a tap, a right-or-wrong every few seconds. That's not repurposing, that's re-engineering for a completely different attention economy.
This is the bit that automated content tools genuinely struggle with. It's easy to build a pipeline that takes an article and produces a video with a voiceover and some stock footage. It's much harder to build one that understands when to interrupt the viewer with a question, when to compress six paragraphs into four seconds of on-screen text, and when the source material simply doesn't deserve to survive the transition. That's editorial judgement, not transformation.
Where Automation Actually Wins
None of that means the students have found something machines can't touch — it means they've found the bit machines need to get better at learning, not skip.
Where bulk automation genuinely beats a small team doing this by hand is volume and consistency. ScrollEd, by the nature of being a handful of students, can only process so many textbooks so fast. If you're a publisher, a training provider, or a marketing agency sitting on ten thousand articles or a back catalogue of case studies, you don't want a boutique studio approach — you want something that can chew through the lot overnight and apply the same quality bar to segment three thousand as it did to segment one.
That's the gap tools like mine are built to close. Not to replace the judgement of "should this become a quiz question" but to make the mechanical parts — segmentation, pacing, voice, format conversion — fast enough that a human only needs to steer, not do the whole job by hand.
What Founders Should Actually Steal
If you're building anything content-adjacent, the lesson from ScrollEd isn't "add quizzes" — it's narrower and more useful than that.
Look at content you already own that nobody's engaging with. Not content you need to create from scratch — content sitting dormant, unread, doing no work. Then ask what the native format of your audience's attention actually is, and stop assuming your original format deserves to survive the trip. A textbook chapter didn't need to become a slightly shorter textbook chapter. It needed to become something built for the scroll.
That's the repurposing instinct I'd want every founder to have before they even think about which tool automates it. Automation without that instinct just produces content faster in the wrong shape. The students at ScrollEd got the shape right first, by hand, which is exactly why it's worth studying before you go anywhere near a pipeline.
The Real Opportunity
The honest takeaway for anyone running a content business: watch what small, resource-constrained teams do when they're forced to be ruthless about format. They can't afford to be lazy about it — they don't have the budget to brute-force bad content into good engagement. That constraint is the gold. Automate the mechanics once you understand what they've already worked out by necessity, and you'll build something that scales without losing the thing that made it work in the first place.
Live your dream — just make sure the format fits the dreamer.
— Wayne