A deep-product marketing leg to convert high-ticket customers

The hook

Step one: what visual most clearly explains “fabric doesn't look like fabric in AI”? Something involving clothing. The presenter is wearing clothes, so his own outfit could carry the point.

Step two: a quick colour switch on that outfit. It breaks attention without fighting the voiceover, it is new, and it is tied directly to the subject. The first-person framing and the opening on-screen text add to it, so even a viewer outside the target audience is held for the first five to ten seconds.

Every frame after

See-say-sync throughout. Key text is highlighted on screen, screen recordings zoom in on each UI element as it is clicked, and smooth motion graphics carry the transitions. Built on the assumption that the viewer has no reason to stay.

Ends on the “agent” comment trigger into the long-form YouTube tutorial.

The hook

Step one: the clearest visual for this subject is the input and the output, the phone shot and the AI shot it produced. So instead of a verbal setup explaining what the video will show, the hook shows both at once in a half-and-half split screen.

Step two: the text overlay reads “I just got this exact camera movement using nothing but just my phone.” The highest-leverage words, nothing but just my phone, come first.

Every frame after

Because the steps happen in the real world, the body is behind-the-scenes footage of someone actually following the process, which is more watchable than screen recordings for this kind of content.

The before-and-after split comes back every few seconds as a structural device. It works like a game: the viewer sees the before and tries to predict the after. Later examples of other shot types use the same split for the same reason.

The hook

Step one: the opening on-screen text delivers the exact takeaway, which is also the solution to the viewer's problem: using references for AI fashion campaigns.

Step two: highly cinematic visuals that read like a high-end fashion campaign, and a stop-motion effect with the campaign's characters moving from right to left. Cuts and motion graphics make it engaging, and the visuals stay locked to what the text says so the viewer knows instantly what the video is about.

Every frame after

For every sentence of voiceover, the same order. First, find the visual that explains the line most clearly. Then use motion graphics, the way it is visualised, and sound design to make it interesting to watch.

The opening slide

The payoff gets placed on the opening slide: “How to get 100% fabric consistency across your AI product ads.” The visual shows four shots of the same fabric staying consistent across motion, wind, sunlight and touch. Whether someone saves it, is sent it, or scrolls past it, they know the exact outcome.

Slide by slide

Every slide after this is optimised for ease of understanding and preventing cognitive overload. Slide two covers only setting up the agent's context. Slide three covers only how to talk to the agent, like talking to a crew. Each slide after that takes one micro-step with a clear visual.

A receipts slide benchmarks real cost and effort against a traditional photo shoot, the comparison enterprise clients kept asking for. The last slide shows the YouTube tutorial with the “agent” trigger.

The opening slide

The value is stated plainly: “How to get accurate B-roll for any location in less than 2 hours.” The one freedom on this piece was the visual, so a drone shot was chosen to make the opening frame stand out. Elevating that frame was the main creative decision available.

Slide by slide

The first couple of slides establish why this is a need at all, meeting the audience at what they already know before walking them into the solution. Then one micro-step of the Agent One workflow per slide, with clear visuals and little text.

A receipts slide shows what a thirty-second B-roll video costs, making the case that this is faster and cheaper than a shoot. The last slide points to the YouTube tutorial.

The opening slide

The value goes up front immediately: $75, under an hour. The visual is lifted well beyond a screen recording, a striking UGC-style shot of the talent with the product. That kind of image is tempting and stops the thumb, and it signals both the production quality and the exact format being discussed.

Slide by slide

Slide two shows the format being replicated and nothing else, so the viewer can take it in. The problem slide explains what happens when you try it manually: three moving parts that all have to match.

The process slides are pure screen recordings. Between them sit breaker slides that name the problem, the catch or the insight at that point, using the same talent holding the product. They lift the visual without breaking clarity, because they are still on subject. A receipts slide gives cost per ad and time per ad, then the CTA.

The opening slide

It names the target persona outright: “If you're an AI filmmaker who wants to test out AI VFX hybrid workflows,” with “Don't skip this” underneath. That line works as a filter, telling exactly who should watch and why. The visual front-loads the most exciting footage from the shoot, a fight sequence, because this is where the most visual play was possible.

Slide by slide

Slide two is “The Big Question”: which parts of filmmaking have to be shot, and which can AI handle? That is exactly where this audience's uncertainty lived. Then the bottleneck, then a slide saying we took a shot at figuring it out.

Every frame after that is deliberately minimal: one strong hybrid or behind-the-scenes shot with only the text it needs. The CTA offers the files so people can experiment themselves.

The opening slide

The one available leeway was the framing. Instead of front-loading the claim, the post says every viral character consistency workflow on the web was tested, and here is the one that works. That opens a curiosity loop: people want to see which one won. The viewer still gets the same useful outcome. The framing just gives them a reason to swipe to it. A pattern-breaking visual sits on the opening frame.

Slide by slide

Each slide pairs an eye-catching, pattern-breaking visual with very simple text: the stakes, the false promises other character sheets make, or, over screen recordings of the other workflows being tested, a minimal overlay. Simple and clear on the subject, and fun to watch. Micro-steps and the CTA as on every carousel.

The opening slide

“Previz” is insider vocabulary that most people outside film won't recognise. It was used anyway, because a filmmaker who sees it stops: it names a pain they know. Stopping the right person mattered more than being legible to everyone, a deliberate trade-off.

The reward goes up front, “We cracked the best way to previz your film with AI,” over a visual that alternates between the sketched previz and the actual shots from the film. That back-and-forth demonstrates the value in the most immediate way and still stops the thumb.

Slide by slide

Slide two, “Why this matters,” changes the background drastically from frame to frame to catch the eye while staying on subject. Every frame after follows the same rule: the visual has to justify the text and still be engaging. Several workflows are shown, then the CTA to the YouTube tutorial.