Building an AI-Powered Creative Automation System with n8n, fal.ai, and Figma

members
3 min read
Building an AI-Powered Creative Automation System with n8n, fal.ai, and Figma

This n8n workflow automates the creation of marketing assets by combining AI scene generation, image creation, copywriting, and design tools into a single pipeline.

What This Workflow Does

The system takes a product image and brand context as inputs, then automatically generates multiple marketing creatives through three main stages:

Scene Generation: Uses Google Gemini to analyze your brand context and product landing page, then creates 10 different scene descriptions where your product could be featured (like "in a yoga studio" or "on a bench in a mountain scene").

Image Creation: Takes each scene description and your product image, generates detailed prompts, then uses google nano-banana via fal.ai's image generation models to create photorealistic product shots in those environments.

Copywriting: Analyzes the generated images and creates marketing copy (headlines and bullet points) tailored to each scene using GPT-4.

All outputs are stored in Google Sheets and can be synced directly into Figma templates for immediate use in design work.

Technical Architecture

The workflow uses a switch-based routing system that handles four different operations:

  • scenes: Generate scene descriptions
  • images: Create product images for selected scenes
  • copywriting: Write marketing copy for generated images
  • sync: Update Google Sheets with asset information

Data flows through Google Sheets as the central hub, with Google Drive for file storage and Google Docs for brand context. The system includes proper error handling and status checking for the fal.ai image generation process.

Setup Requirements

You'll need authenticated connections to:

  • Google Drive, Sheets, and Docs
  • OpenAI (GPT-4)
  • Google Gemini
  • fal.ai

File Preparation:

  1. Copy the provided Google Sheet template
  2. Create two Google Drive folders: one for generated creatives, one for product assets
  3. Make folders publicly accessible
  4. Create a brand context document using the provided template

Brand Context File: This is crucial. The template asks for specific information including brand overview, target audience, tone of voice, creative guidelines, and example content. The quality of your outputs depends heavily on how well you fill this out.

Workflow Configuration:

  1. Update the Settings node with your sheet and folder URLs
  2. Activate the workflow and copy the webhook URL
  3. Add your product image to the assets folder
  4. Use the sheet's automation menu to sync the folder and select your asset

Running the System

The process is straightforward:

  1. Hit "Generate scenes" in the Google Sheet menu
  2. Review generated scenes and mark which ones you want images for
  3. Run "Create Images" to generate product shots
  4. Run "Write Copy" to create marketing text
  5. Sync everything to your Figma template

Technical Considerations

The image generation process includes proper polling logic since fal.ai runs asynchronously. The system checks job status every 3 seconds until completion.

The copywriting stage passes the actual generated images to the AI agent, not just descriptions, enabling more accurate and contextual copy creation.

Google Sheets serves as both the control panel and data store, making it easy to review outputs, make selections, and track what's been generated.

Limitations and Requirements

This isn't a simple plug-and-play solution. You need:

  • Multiple API accounts and authentications
  • Understanding of n8n workflow management
  • Time to properly configure your brand context
  • Figma knowledge to set up templates

The system works best when you invest time in creating detailed brand guidelines and scene descriptions. Generic inputs will produce generic outputs.

The workflow handles errors reasonably well, but failed image generations or API issues will require manual intervention to retry specific items.

This represents a practical approach to creative automation - not replacing human creativity, but handling the mechanical aspects of generating multiple variations once you've defined the creative direction.

How-To Document and n8n Workflow-Download

Find the downloads below: