How to Create a Motion Graphic Ad with AI: A Step-by-Step Workflow

Learn how to create a motion graphic ad with AI using Claude Code, ElevenLabs, Audacity, and a clear workflow for character design, voiceover, animation, and review.

You can create a short motion graphic ad with a small set of AI and editing tools—if you plan the visuals, script, and voice before assembling the video. This guide walks through a workflow for making a character-led ad, from preparing a reusable motion-graphics skill to reviewing the final animation.

The example is based on a creator’s workflow for a short promotional video in Egyptian Arabic. It uses Claude Code to help plan and build the video, ElevenLabs for text-to-speech, and Audacity to edit the voiceover. The same approach can be adapted to other languages, characters, and products.

Note: Results depend on your tools, input files, and editing choices. Treat AI output as a first draft that needs review.

1. Plan the ad before generating it

Start with a short creative brief. Write down:

What product or app the ad promotes
Who the audience is
The main message
The language and dialect
The desired tone, such as playful, calm, or energetic
The approximate video length and format
Any brand colors, fonts, or character references

A clear brief helps keep the script, character, and scenes consistent. If the ad features a recurring character, prepare a character sheet before asking an AI tool to create multiple scenes. Include the character’s clothing, hairstyle, age range, key features, expressions, and the angles you need.

For a real campaign, use original references or assets you have permission to use. A style reference can help communicate pacing and mood, but copying a creator’s exact scenes, character, or distinctive design can create intellectual-property problems.

2. Use a repeatable workflow in Claude Code

The attached workflow describes using a custom motion-reel skill with Claude Code. A skill is a reusable set of instructions that can guide an AI through a repeated process. Anthropic’s documentation explains how skills can be used and organized in Claude Code. Anthropic’s Agent Skills documentation

In this example, the skill acts like a production checklist. It tells Claude Code to ask for missing information—such as the video length, language, voiceover, references, and visual identity—before planning the scenes. It can also set rules for pacing and transitions.

A useful skill brief might specify:

Keep each scene short and focused on one idea.
Change the visual every few seconds when it supports the narration.
Use motion that fits the character and tone.
Keep captions readable and make sure they match the spoken language.
Preview the storyboard or scene plan before rendering.
Review the rendered result and revise obvious timing or visual issues.

A skill can support a workflow, but it does not guarantee that every scene will be correct. Check the plan and output yourself, especially for spelling, language direction, character consistency, and timing.

3. Create a visual identity and character sheet

Before producing the scenes, settle on a small visual system:

A limited color palette
One or two typefaces
A logo or placeholder for the product
A consistent background style
A character sheet for any recurring person or mascot

The PDF’s example used an older Egyptian woman as its central character and created a separate character sheet so that the character could remain consistent across the ad. That is a useful production step: a reference sheet gives the video tool a stable visual guide, though generated images may still vary between scenes.

Check that text is legible on mobile screens. For Arabic, verify right-to-left layout, letter connections, and the exact wording in the final video. Do not rely on a generated frame alone to confirm correct Arabic typography.

4. Write and test the voiceover script

Write the script for the ear, not just the page. Use short sentences, natural phrasing, and pauses that fit the intended performance. For a dialect such as Egyptian Arabic, read it aloud or ask a fluent speaker to review pronunciation and tone.

The example workflow generated a voiceover with ElevenLabs, then tried several versions before choosing a preferred take. ElevenLabs’ text-to-speech documentation describes voice selection and ways to guide delivery with text context and supported audio tags. ElevenLabs Text to Speech documentation

To improve pronunciation:

Listen to the first generated version.
Mark words the voice mispronounces.
Rewrite or phonetically clarify those words where appropriate.
Adjust punctuation and performance cues.
Generate a few alternatives and compare the complete takes.

Do not assume a voice will pronounce every dialect or name correctly on the first try. Also check the current service terms and plan details before using generated audio in a commercial ad.

5. Edit and choose the best voiceover take

If the generated narration needs cleanup, use an audio editor to compare takes and assemble the strongest performance. In the PDF’s example, the creator used Audacity to trim and combine preferred sections from several generated versions. Audacity is an audio-editing application; see its official download page and installation guide.

When editing:

Keep the narration clear and at a consistent volume.
Remove awkward pauses and unwanted noise.
Use music and sound effects carefully so they do not compete with speech.
Listen to the full edit from beginning to end.
Confirm that the voiceover still sounds natural after combining sections.

Make sure you have the rights needed for any music, sound effects, voices, and other assets used in a commercial project.

6. Build the scenes around the narration

Once the voiceover is ready, use it to shape the scene timing. Each scene should communicate one visual idea that supports the spoken line. Keep the character, color palette, and typography consistent from shot to shot.

A practical sequence might look like this:

Opening hook: Introduce a relatable problem or question.
Product reveal: Show how the app or service helps.
Key benefit: Demonstrate one or two useful features.
Closing: End with a clear call to action.

Use transitions and motion to clarify the story. Effects should not distract from the message. If a scene includes a recognizable third-party logo or product, consider whether you have permission to use it and whether the scene could confuse viewers about who made or endorsed the ad.

7. Review, revise, and export

Before exporting, review the video more than once. The creator’s workflow in the PDF includes inspecting rendered frames and making revisions before considering the ad finished. A structured review helps catch issues that are easy to miss while generating scenes.

Use this checklist:

Does the opening make the product and audience clear?
Does the voiceover match the scene timing?
Are character details consistent?
Is every caption spelled correctly and readable?
Does Arabic text display right-to-left correctly?
Are music and sound effects balanced against speech?
Are the logo, product images, and references used appropriately?
Does the final call to action make sense?
Does the exported video play correctly on mobile?

If you use FFmpeg as part of your rendering or conversion workflow, consult the official FFmpeg documentation for the options relevant to your format and export settings.

Example prompt for a motion-graphics workflow

Use this as a starting brief for your project, then replace the bracketed details:

Help me create a short motion graphic advertisement for [product or app]. The target audience is [audience], and the main benefit is [benefit]. The video should be about [duration] long and use [language and dialect] with a [tone] tone.

First, review the attached brand identity, character sheet, voiceover, and visual references. Ask me for any missing details. Then prepare a scene-by-scene plan with timing, narration, on-screen text, character actions, transitions, and sound notes. Keep the character and visual identity consistent. Use references to understand pacing and mood, but create original scenes and do not copy protected logos, artwork, or exact designs. Show me the plan for review before rendering. After rendering, identify issues in the scenes and suggest specific revisions.

Final thoughts

A strong AI-assisted motion graphic ad depends on preparation and review as much as on generation. Define the brief, establish a visual identity, test the voiceover, build scenes around the narration, and inspect the final export. AI tools can speed up parts of production, but human review is what helps make the ad accurate, coherent, and ready for its audience.

Sources and tools

Arabic Guide PDF Here