Reimagine marketing at Volkswagen Group with generative AI

1 views Source
Reimagine marketing at Volkswagen Group with generative AI

Volkswagen Group, one of the world's largest automotive manufacturers, collaborated with the AWS Generative AI Innovation Center in 2025 to develop a solution harnessing generative AI. This solution enables the creation and evaluation of marketing images of vehicles while maintaining brand precision. The result is an end-to-end marketing image generation and evaluation pipeline utilizing image generation models hosted on Amazon SageMaker and image evaluation powered by Amazon Bedrock.

The marketing teams at Volkswagen faced the challenge of producing thousands of marketing assets annually while ensuring that each image met strict brand standards. A single vehicle launch might require hundreds of image variations, traditionally taking months of production work. On-location photo shoots can cost six figures, requiring physical prototypes and complex logistics. However, if Volkswagen could generate photorealistic images in minutes instead of weeks, it would open new opportunities for cost reduction and faster time-to-market.

To achieve this goal, it was essential to generate images indistinguishable from professional photography and aligned with the visual standards of each of the Group's ten brands. Initial experiments with base models revealed that they did not account for Volkswagen's unique design language, limiting their applicability for marketing campaigns.

The solution involved fine-tuning models on Volkswagen's proprietary visual assets. The team employed DreamBooth techniques for fine-tuning on images, allowing for the creation of high-quality data with precise control over vehicle specifications and environmental conditions. The model was deployed on Amazon SageMaker AI, ensuring efficient processing and scalability.

However, generating images successfully required the right prompts. The team discovered that effective prompts for automotive marketing imagery necessitated specialized vocabulary and style, often lacking among users. This highlights the importance of accuracy and specificity in the image generation process to meet the high standards of the brand.

Related articles