Bringing product accuracy to generative AI

Accelerated by NVIDIA technology, ProductDrop helps consumer brands create campaign and e-commerce images with AI while keeping products accurate across entire ranges and markets. Read the press release.

Omi is launching ProductDrop to combine AI-generated environments with its Digital Twins: approved 3D replicas of real products that preserve their shape, materials, packaging and artwork.

A beauty brand can, for example, place its shampoo range in an AI-generated bathroom, then apply that creative concept across sizes, packaging versions and markets. Each image uses the appropriate approved product, giving teams the freedom to explore new settings while keeping their products accurate.

Why product accuracy matters at scale

Generative AI has expanded what creative teams can imagine. But when an AI tool generates a product, it can change details that matter: the shape of a bottle, the color of a lipstick or the wording on a label. An image may look convincing while showing a product that does not match what customers can buy.

For brands managing hundreds or thousands of products, accuracy must hold across an entire catalog. A single campaign may span dozens of shades, several pack sizes and different market versions.

Those products also keep changing. Packaging is redesigned, claims are updated and new variants enter the range. Brands need an approved product reference that stays current and a way to carry those changes through to their content.

Omi’s Digital Twins provide that product source of truth, giving teams a maintained, reusable representation of each product across campaigns, markets and creative tools.

“Your product is at the heart of your brand. You can’t compromise on how it looks. AI gives brands the freedom to create any world around it. Digital Twins make sure the product entering that world remains an accurate, approved representation of the real product. With ProductDrop, that creative freedom can finally scale across thousands of products.”

Hugo Borensztein

CEO & Co-Founder, Omi

How ProductDrop works

Teams start with an approved Digital Twin and an AI-generated or existing image. ProductDrop turns that image into an editable 3D environment, allowing teams to control product placement, camera angles, lighting, shadows, reflections and transparency.

This helps the product belong in the scene. A glass bottle can reflect its surroundings and let the environment show through it. A pack can cast a shadow that matches the surface beneath it. The Digital Twin preserves the product’s geometry, materials and artwork throughout the process.

Once the setting is established, teams can reuse it across a product range, bringing the appropriate Twin into the scene for each shade, size or market. They can also use their approved product catalog with environments created through different AI tools.

When a product changes, teams update its Digital Twin and regenerate affected visuals from the current approved version. A packaging update can carry through to campaign and retailer content, reducing the need to correct individual images.

How L’Oréal is using this approach

L’Oréal is already combining Omi’s Digital Twins with AI-generated brand environments within its CreAItech program.

Teams combine Omi’s Digital Twins with AI-generated brand environments. The group reports producing high-quality renders in just two days, enabling local teams to adapt content independently across markets. Marketers can apply each brand’s creative environments across product lines without specialist design expertise.

“We use Omi to combine accurate 3D Digital Twins of our products with the backgrounds and creative environments we generate using AI. This gives our brands greater creative freedom while preserving the accuracy of our products, without GenAI reproduction errors.”

Thomas Alves Machado

GenAI Content Global Director, L'Oréal Groupe

NVIDIA acceleration reduces production time

Combining AI-generated scenes with realistic 3D Digital Twins requires substantial computing power. NVIDIA technology supports the scene analysis, image generation and rendering behind ProductDrop, helping teams move from creative exploration to production across large product ranges.

ProductDrop is built on NVIDIA OptiX and NVIDIA CUDA, combined with open-source image understanding and generation models including the FLUX family and Qwen-Image, running on NVIDIA RTX PRO 6000 Nvidia Blackwell Server Edition GPUs. Omi’s pipeline also integrates open-source models such as MoGe for depth estimation and Microsoft TRELLIS for 3D asset generation.

"For retail and CPG brands, product accuracy isn’t optional — it’s the foundation of consumer trust. NVIDIA’s accelerated platform, built on CUDA, OptiX, and RTX PRO 6000 Blackwell GPUs, was designed to bring exactly this kind of precision and scale together. What Omi has built with ProductDrop shows how open-source innovation and purpose-built infrastructure can transform creative production for the industry"

Azita Martin

VP and GM of Retail and CPG, NVIDIA

ProductDrop is served through a custom inference endpoint on Verda, a European full-stack AI cloud, TRELLIS inference runs 10x faster, cutting overall generation pipeline latency by 3x.

“Our AI Lab’s work with Omi shows what’s possible when inference serving is optimized specifically for a workload, in this case cutting TRELLIS latency by 10x. Speed shouldn’t come at the cost of quality, and that’s exactly the kind of problem our applied research is focused on solving,”said Paul Chang, AI Infrastructure Lead at Verda.

Press enquiries

Read the press release. For an interview with Omi’s founders, a ProductDrop demonstration or high-resolution images showing the workflow, contact Agence Ballou at omi@balloupr.com.

About Omi

Omi is the Product Source of Truth for consumer brands. Its platform generates product photos, videos and AI campaign visuals from approved 3D Digital Twins. Teams can scale content across entire catalogs and regenerate affected visuals when packaging, variants or market requirements change.