To cut a 30-second teaser for an AI short film, choose one story question, build a short sequence around it, and stop before the answer. Give viewers a clear route to the complete film. Start with the finished edit: selecting and rearranging existing material is usually a better first step than generating more shots.

Prompt2Screen’s Torino collection brings together ten films made in under ten hours from an Automotive brief revealed that morning. The event took place at Blox Space on 10 October 2026. Once a film is finished, a new editing problem begins: how do you introduce it to someone who was not at the screening?

The workflow below is an editorial exercise, not a Prompt2Screen submission requirement or a formula for audience growth.

Decide what the teaser should make someone wonder

Write a question that belongs to your film. “Will she leave?” gives you something to build toward. “How good does this look?” tends to produce a collection of impressive shots without a reason to keep watching.

Choose one audience action as well: watch the full film, explore a collection, or read about the process. A single-film teaser and an event sizzle have different jobs. The first introduces one story; the second can introduce the range of work made by several teams. Avoid trying to accomplish both in the same thirty seconds.

For this exercise, imagine a fictional automotive short about a daughter deciding whether to restore her father’s abandoned car. This is an invented example, not a description of a Torino entry. The teaser question could be: “Will she bring it back to life?”

Select moments by their job in the edit

Duplicate your finished timeline and make a selects sequence. Find a moment that establishes the person or situation, another that introduces resistance, and one that makes the unresolved question feel worth answering.

For the invented film, those might be a hand removing a dusty cover, an ignition that fails, and the daughter looking at a photograph on the dashboard. The most elaborate generated camera move may contribute less than a small reaction.

Keep enough of each shot to try different cuts. Listen before removing its sound: a breath or mechanical click may create a stronger connection than a title card. Do not assemble unrelated dialogue fragments into a statement the film never makes.

Try this 30-second structure, then adjust it

Use these timings as a starting edit, not a rigid template:

  • 0–4 seconds — a specific opening: the cover comes off the car; we see the daughter’s reaction.
  • 4–11 seconds — enough context: the photograph and a detail of the neglected interior establish a personal connection.
  • 11–20 seconds — resistance: the ignition fails; she pauses, then reaches for a tool.
  • 20–25 seconds — an unresolved turn: a sound or movement suggests something might change. Cut before showing whether the repair succeeds.
  • 25–30 seconds — title and next step: show the film title, a concise creator credit and “Watch the full film,” with the destination provided alongside the post.

Watch the sequence without explaining it. If viewers cannot tell what she wants, give context more time. If they already know the outcome, remove the revealing shot. If the title cannot be read comfortably, shorten the montage instead of shrinking the type.

A slow, observational film may need fewer cuts and longer holds. Preserve its character: a frantic teaser can set expectations the complete film will not meet.

Make the sound transition feel intentional

Carry a useful sound across a picture cut to connect two moments. In the example, the ignition attempt could start before the close-up of the key. A brief silence after the failed start may give the next image more weight than another musical hit.

Listen for abrupt room-tone changes, dialogue clipped mid-word and a music ending that feels accidental. Keep the sound level comfortable through the final title. For a more detailed review, use our five-pass sound-design workflow.

Check dialogue captions against this new edit. Captions copied from the full-film timeline can have the wrong timing after shots move. Our captioning guide walks through reviewing the text against the exported video.

Reframe deliberately for each placement

Make a separate vertical edit if you need one. Inspect every shot: an automatic centre crop can remove the photograph, the other character or the direction of a glance. If a wide composition carries essential information, choose a different shot or design a layout that keeps it readable.

Preview the finished version on a phone with the destination platform’s interface visible. Keep names and captions away from controls that cover the picture. Test whether the central situation is understandable with the sound muted, while still treating sound as part of the intended experience.

For a standard YouTube video, its thumbnail guidance recommends a 16:9 image. Choose a frame that represents the film honestly and remains clear at a small size. Avoid putting a visual reveal in the thumbnail that you deliberately withheld from the teaser.

Give viewers a useful next step

Link directly to the full film or the relevant collection, and check that the destination is public and works on mobile. Name the film and its creators in the description. A viewer should not have to search through an unrelated homepage to find the ending.

If the follow-up is about how the film was made, prepare a separate process note. VIA Group’s Milan AI Festival has a useful guide to building an AI film project breakdown: it focuses on creative decisions and supporting examples. Keep that explanation distinct from the story teaser so each has room to work.

Review evidence before making another version

Ask a few viewers what they think the film is about and whether they know where to watch it. Their answers can expose unclear context or a missing link before you start comparing performance numbers.

On YouTube, review retention when it is available. YouTube’s retention documentation says data typically needs one to two days to process. It also specifies that highlighted key moments require a video of at least sixty seconds and at least one hundred views, so do not expect those highlights on a thirty-second teaser.

If you test another cut, change one meaningful decision: the opening shot, the amount of context or the final invitation. Record what changed and compare similar traffic sources where possible. A small difference in views alone does not establish which edit tells the story better.

Watch the Torino films for different approaches to introducing a story. Aim to leave viewers curious, oriented and one clear step away from the full film.

Sources & further reading

Updated 16 October 2026. Event availability and final details are maintained on the linked official event pages.

Your next scene starts here.

Torino was the beginning. Register your interest in the next edition. More will be revealed on 15 November 2026.

Register your interest ↗