This is a production problem, and it shows up in the data before most teams know what to call it.

In the Inceptly Modular Creative System, every ad is built from three segments: the Intro, the Bridge, and the Core. Each segment has a specific job — the Intro earns attention, the Bridge shifts belief, the Core earns the click.

The Intro is where visual variation matters most, and where most teams are doing the least of it.

An Intro isn't just a hook line. It's a visual format with a specific psychological mechanism. An AI UGC hook, a podcast clip, a text-on-screen opening, a pattern interrupt, a split screen — each one carries a different attention signal and qualifies a different type of viewer in a different psychological state. They're not stylistic choices. They're different tools doing different jobs.

When a team produces 50 ads and 48 of them open with the same AI UGC character, they've tested one format 48 times with slightly different copy. The creative learning from that is almost zero.

Real variation means testing different Intro formats, not just different scripts inside the same one.

These are the primary hook formats used across Inceptly campaigns.

1. Pattern interrupt. Unexpected, native-feeling, unrelated to the product. Its only job is to stop the scroll. The Bridge does the qualification — the interrupt doesn't need to connect logically to what follows. A common mistake is rejecting strong interrupt concepts because they feel "off-brand." That's the wrong filter.

2. AI podcast. Two people in a studio environment. Audiences associate this format with expert-level content, which transfers a credibility signal before a single product claim is made. The format to reach for when the Bridge depends on mechanism education or authority positioning.

3. AI UGC. Someone speaking to the camera in a casual setting. The strongest authenticity signal available in an AI production workflow. Works best when the Bridge relies on permission or social proof — "someone like me already did this." The character should be built to match the target audience profile, not a generic creator archetype. At Inceptly, we've shifted toward AI UGC over real UGC almost entirely for speed, and the aesthetic we're after is deliberately unpolished. Slightly dirty lens, imperfect framing. The more it looks produced, the less it performs. Audiences have calibrated to this faster than most production teams have.

4. AI street interview. A presenter in a public outdoor environment, man-on-the-street style. Creates an impression of public consensus. The right opener for Bridges that need the viewer to believe the topic is broadly relevant rather than niche.

5. Text on screen. Bold typography, no face, no character. Underused, especially by teams that have convinced themselves production complexity signals quality. When the copy is strong, this format performs comparably to much higher-cost executions. Clean, fast to produce, and effective in feeds that have become saturated with talking heads.

6. Split screen. Two states, two outcomes, two approaches shown simultaneously. The visual structure does the framing work before any copy lands. Most useful for contrast-based setups where the ad's central tension needs to be established in the first two seconds. One production note: on vertical video formats — Shorts, Reels, TikTok — always split horizontally (top and bottom panels). A vertical split on a vertical frame compresses each panel to a width that isn't legible.

7. Motion graphic intro. Animated text, stylised design, no live footage. Increasingly important as AI-generated content raises the baseline quality of everything in the feed. A motion graphic intro signals "this is different" before a word is spoken, and because it's a self-contained, interchangeable unit, a single well-produced animation can be tested against multiple Bridges and Cores without rebuilding anything.

Once you have a working Bridge and Core, those components don't need to change to test whether your hook format is working. You swap the Intro only.

One Bridge, one Core, tested against seven different Intro formats. Seven genuinely distinct signals to the algorithm. Seven real data points about what stops the scroll for this audience — more useful creative learning than most teams generate from fifty ads.

The tooling to make this feasible keeps getting easier. AI video generation has moved fast enough that producing a podcast-format hook, a street interview sequence, or a motion graphic intro is now primarily a question of knowing what you want, not whether you can make it. On the production side, editors at Inceptly now work with AI-built workflow helpers — custom bots within LLMs that generate prompts, shot descriptions, and production briefs — which compresses the time between concept and output significantly.

When every ad opens the same way, you end up optimising inside a single format rather than finding which format works. The budget goes to the best-performing UGC hook out of forty UGC hooks, when a motion graphic intro or a text-on-screen opening might have outperformed the entire batch at a fraction of the production cost.

The teams seeing the most useful data right now are the ones asking "what format should the hook be" before they ask "what should the hook say." Format is the variable most production teams are leaving on the table.

Volume matters. But volume inside a single visual format is just expensive noise.

Miro and the Inceptly Team

Get in touch with us by responding to this email or tagging us on LinkedIn or Instagram, and sharing your thoughts. Your feedback helps us keep our newsletter relevant and interesting.