Fashion AI SchoolLab

intermediate · 6 steps

From Idea to Campaign: Image, Video, Edit, Launch

Turn a single campaign idea into a fully produced, multi-format ad launch using a chain of AI tools for research, imagery, motion, and editing.

  1. Perplexity
  2. ChatGPT
  3. Midjourney
  4. Runway
  5. Kling
  6. Photoshop (Generative Fill)

Goal

Take a raw campaign concept and produce a polished, launch-ready set of images and video assets in a single afternoon.

Expected result

A complete, on-brand campaign package: hero images, a 15-20 second video, and platform-ready resized assets, produced end-to-end in one working session.

  1. 1
    Perplexity

    Research the campaign angle

    Before generating anything, ground the campaign in real market context. Use Perplexity to pull recent trend reports, competitor campaign examples, and audience sentiment around your product category. This step prevents you from producing beautiful but irrelevant content. Save the three strongest insights and one surprising stat to reference later in your creative brief and copy.

    Prompt for Perplexity
    Research current marketing campaigns and consumer sentiment for [PRODUCT CATEGORY] aimed at [TARGET AUDIENCE]. Summarize the top 3 creative trends, 2 competitor campaign examples with what worked, and one surprising statistic I can use in ad copy. Keep it concise and cite sources.
  2. 2
    ChatGPT

    Write the creative brief

    Convert your research into a tight creative brief that defines the campaign message, tone, hero visual concept, and channel mix. This becomes the single source of truth the rest of the team (and every AI tool downstream) will follow, keeping outputs consistent across images, video, and copy.

    Prompt for ChatGPT
    Using this research: [PASTE RESEARCH SUMMARY], write a one-page creative brief for a campaign promoting [PRODUCT NAME]. Include: campaign message, tone of voice, target audience, hero visual concept, 3 image scene ideas, and a 15-second video concept.
  3. 3
    Midjourney

    Generate hero campaign imagery

    Turn the hero visual concept from your brief into a set of striking campaign images. Generate multiple variations of the same scene to give yourself options for cropping into different ad formats (square, story, banner). Focus on lighting and composition that leaves room for text overlay later.

    Prompt for Midjourney
    Editorial fashion campaign photo of [MODEL DESCRIPTION] wearing [PRODUCT DESCRIPTION], [SETTING/LOCATION], cinematic lighting, shallow depth of field, bold color palette matching [BRAND COLORS], negative space for text overlay --ar 4:5 --v 6
  4. 4
    Runway

    Animate the hero image into video

    Bring your chosen still image to life with subtle motion — fabric movement, hair drift, camera push-in — to create a short looping video for social ads. Keep the motion understated so it reads as premium rather than gimmicky, and export at a length that suits Reels/TikTok pacing.

    Prompt for Runway
    Animate this image with a slow cinematic push-in, gentle fabric and hair movement in the wind, subtle depth parallax, keep the subject sharp and in focus, 4 second loop, no camera shake.
  5. 5
    Kling

    Extend into a short campaign video

    Turn your animated clip into a longer narrative sequence by generating a follow-up shot that continues the same scene, such as a turn, a walk, or a product close-up. Stitch this with your Runway clip to build a 15-20 second campaign video with a clear beginning and end.

    Prompt for Kling
    Continue this scene: the model turns slightly toward camera and smiles softly, then the shot transitions to a close-up of [PRODUCT DETAIL], maintaining consistent lighting and color grade, 6 second clip, smooth camera motion.
  6. 6
    Photoshop (Generative Fill)

    Retouch, resize, and launch

    Clean up any AI artifacts, extend backgrounds to fit each platform's aspect ratio using generative fill, and export the full asset set. Then assemble the final ad copy, images, and video into channel-ready formats in Canva before scheduling the launch.

    Prompt for Photoshop (Generative Fill)
    Use generative fill to extend the background of this campaign image to a 9:16 vertical crop while preserving the subject and lighting, remove any visible artifacts on hands and fabric edges, keep skin tones natural.