Skip to case study
EP
Ekaterina PushkinaMotion Designer

homeOSS4AI

Long-Form to Short-Form
Motion System

A reusable content-production workflow that turns long-form educational material into focused short-form campaigns. Its first use transformed a two-hour lecture on AI agents into 15 standalone videos—each designed to teach one useful idea, build instructor credibility, and lead relevant viewers to the Maven course.

Short-form content strategyAI-assisted editorialPremiere ProAfter EffectsRemotionHuman-in-the-loop
View more videos
Subject
Yujian Tang’s AI Agents lecture
Goal
Create useful standalone clips that build trust and direct interested viewers to the relevant Maven course
My role
Editorial strategy, motion-system design, AI production direction, final polish and approval
Challenge 01

An interesting moment may not work as a complete short-form clip.

View full challengeHide details

Yujian's lecture is full of definitions, caveats, and asides that depend on earlier context, and keyword search can find a relevant topic but can't tell whether that section stands on its own.

The evaluation therefore measured audience demand and editorial completeness separately. A clip qualified only if it introduced the topic, explained it, established why it mattered, and ended on a complete thought. A popular topic was rejected if the excerpt lacked context, changed the meaning of a claim, or ended before the main point.

Challenge 02

Static slides could not support a scalable short-form campaign.

View full challengeHide details

Yujian's lecture leaned on static presentation slides, so extracted clips lacked the visual movement short-form platforms need, and building custom graphics for each video by hand would have slowed production.

This created a second requirement: a reusable motion system that could turn each clip's meaning into relevant supporting graphics while following consistent visual, layout, and animation rules.

The workflow automates repeatable tasks while keeping final decisions with the designer.

Local scripts handle transcription and media processing. AI supports research and clip analysis, proposes supporting graphics under designer-defined rules, and implements approved visuals in Remotion. After automated QA, the designer completes the Adobe polish and sound mix, then reviews and approves every delivery master.

Repeatable workflow topology
Pan to explore · click a stage for details
01 Transcribe02 Research + select03 Design + build04 Validate + finish
Stage 01

Candidate score

Topic demand25%
Usefulness + clarity20%
Audience relevance15%
Standalone meaning10%
Hook + retention10%
Visual potential10%
Maven alignment10%
Stage 02

Completeness gate

01Starts without missing context?
02Communicates one clear idea?
03Delivers the promised takeaway?
04Ends on a complete thought?
Assessment categories
Ready to editWorkable with a text setupWorkable with a factual text setupReject: cannot be made self-contained

Selected work

Supporting graphics turn technical ideas into clear visual explanations.

Before
After
Project outcome

The project produced 15 reels and a workflow that can be reused for similar lectures.

The same editing rules, layouts, and Remotion components worked across several AI topics while preserving the speaker’s intended meaning.

Review the workflow ↑