Ulta Beauty has launched a new virtual experience, powered by NVIDIA’s StyleGAN2 generative AI model, allowing users to try on different hair styles and colors. Customers can upload a headshot or use a model’s photo to experiment with various looks, with direct links to Ulta’s products to achieve the styles in real life.
The feature has proven successful, with Ulta reporting that users who engage with the virtual tool are more likely to make purchases. “Shoppers need to try out styles before buying,” says Juan Cardelino, Director of Digital Innovation at Ulta. “Offering virtual try-ons is part of our DNA, whether in-store or online.”
The GLAMlab platform is Ulta Beauty’s first application of generative AI, utilizing NVIDIA’s StyleGAN2, a neural network model known for generating realistic images in multiple styles. “StyleGAN2 was the right choice due to its open-source nature, and we customized it to focus only on hair while preserving facial features,” Cardelino adds.
The tool, available on Ulta’s website and mobile app, uses NVIDIA Tensor Core GPUs for fast AI processing. The first style takes about five seconds to generate, with subsequent looks appearing in less than one second.
Looking forward, Ulta plans to expand the feature to include wigs and explore its integration with in-store styling services. “Stylists could use the tool to show customers how new styles might look, helping them step out of their comfort zones without committing to a permanent change,” Cardelino says.
Related topics:
- Ecoslay Launches Lemon Drop Hair Foam for Long-Lasting Style
- Angelina Jolie’s Stunning Transformation into Maria Callas: Makeup and Hair in ‘Maria’
- Viviscal Hair Growth Supplements: A Proven Solution for Thicker, Stronger Hair
