The real innovation is not image generation. It is a reusable system for producing accurate product content at scale.
Most conversations about generative AI in marketing focus on one question:
How can we create more content?
Nestlé is addressing a more fundamental challenge:
How can a global company build a content system that produces thousands of accurate, localized and channel-ready assets without rebuilding every campaign from scratch?
In June 2025, Nestlé announced an in-house, AI-powered content service for brands including Purina, Nescafé Dolce Gusto and Nespresso. The service uses 3D digital twins—virtual replicas of physical products—to create content for e-commerce and digital media.
At first glance, this might look like a faster alternative to product photography.
The strategic significance is much bigger.
Nestlé is beginning to transform product content from a series of individual creative projects into reusable marketing infrastructure.
What Nestlé actually built
A digital twin is a detailed 3D representation of a physical product. In Nestlé’s case, that could be a coffee machine, a pet-food package or a box of coffee capsules.
Once the digital product has been created, teams can reuse it in different environments and formats. They can change the camera angle, background, lighting, language, packaging variation or campaign setting without arranging an entirely new photoshoot.
The service was developed with Accenture Song, built using NVIDIA Omniverse and OpenUSD, and hosted on Microsoft’s infrastructure. It also uses NVIDIA AI Enterprise for generative-AI capabilities.
This combination provides Nestlé with three important layers:
- An accurate digital representation of the product
- A reusable environment for adapting and rendering that product
- AI-supported tools for producing and scaling creative variations
OpenUSD is particularly relevant because it provides a common framework for organizing, combining and editing complex 3D assets across different applications. It also supports non-destructive changes, allowing teams to modify parts of a scene without permanently changing the original master asset.
That makes the digital twin more than a realistic picture. It becomes a controlled digital source from which many assets can be created.
Nestlé’s official announcement describes the twins as exact 3D replicas that can be digitally adjusted or localized for seasonal campaigns, different markets and channel-specific formats.
Why the traditional content model is struggling
Traditional product-content production was largely designed around campaigns.
A company planned a campaign, organized a photoshoot, edited the selected images and distributed a limited number of finished assets.
Digital marketing has changed that model.
A single campaign may now require different assets for:
- E-commerce product pages
- Retail-media networks
- Instagram, Facebook, TikTok and YouTube
- Display advertising
- Streaming platforms
- Email marketing
- Regional websites
- Multiple languages and packaging versions
Nestlé says campaigns across social media and streaming platforms often require six or more advertising formats. When those requirements are multiplied across products, countries, languages, retailers and seasons, the number of necessary variations grows quickly.
The bottleneck is no longer simply creating the main campaign idea.
The bottleneck is adapting, approving and distributing that idea across every required market and channel.
From one photoshoot to a reusable product asset
The economics of a digital twin are fundamentally different from those of a traditional photoshoot.
A photoshoot usually produces a fixed collection of images. When the packaging, market, format or creative direction changes, at least part of the production process may need to be repeated.
A digital twin requires an initial investment to create and validate the 3D product. But once that master exists, it can potentially support many future assets.
The process moves from:
Physical product → photoshoot → finished image
To:
Verified digital product → controlled variations → multiple finished assets
This does not mean photography disappears. Physical shoots may remain important for human storytelling, lifestyle imagery and campaigns where authenticity is essential.
Digital twins are especially valuable for the repetitive product-production layer: packshots, product angles, localized packaging, seasonal settings and multiple channel formats.
The more frequently a product must be adapted, the stronger the economic case for creating a reusable digital master.
The 70% claim requires context
At the service’s launch, Nestlé reported that it already had a baseline of 4,000 3D digital master products, mainly for global brands. The company stated an ambition to convert 10,000 products into digital twins within the following two years.
Nestlé also reported a reduction of more than 70% in the time and cost associated with scaling digital twins.
This is an important distinction.
The company did not say that every campaign, photoshoot or part of its complete marketing budget became 70% cheaper. The figure specifically concerns the process of scaling its digital-twin capability.
Microsoft later presented the same figure in its analysis of Nestlé’s content system, linking the improvement to faster content production and easier seasonal and channel-specific updates.
The number is still significant. But its real meaning is that Nestlé believes it has found a more efficient way to industrialize the creation and use of digital product masters—not that AI has eliminated 70% of all creative costs.
The organizational model matters as much as the technology
Nestlé’s system is not being introduced as an isolated creative tool.
The company connected it to an existing global content organization that includes 250 marketing specialists across seven marketing hubs and 45 content studios worldwide.
That organizational structure is one of the most important parts of the use case.
A digital twin has limited value if it remains inside a specialist 3D department. Its value increases when local and global teams can access it, adapt it and use it within established approval and distribution processes.
The operating model appears designed to combine:
- Centralized product accuracy
- Shared technical standards
- Global brand consistency
- Local language and packaging adaptation
- Distributed creative execution
This is where AI moves from experimentation into operations.
The objective is no longer to demonstrate that a model can generate an impressive image. The objective is to make the capability available repeatedly across real teams, brands and markets.
Nestlé’s 2025 annual report subsequently confirmed that the company had launched the digital-twin service as part of its wider effort to increase brand-building efficiency.
Why product accuracy is strategically important
General-purpose image generators are powerful, but they can alter details, invent visual elements or reproduce packaging incorrectly.
That is inconvenient for casual content. For a global consumer-products company, it can create serious problems.
Product visuals may contain:
- Logos and protected brand elements
- Ingredients and nutritional claims
- Legally required information
- Market-specific labels
- Exact colors and packaging dimensions
- Promotional statements
- Retailer-specific product information
A verified digital twin can anchor the product itself in an approved source asset. Generative AI can then help create environments and variations around that controlled representation.
This approach does not remove the need for human review. It gives the review process a more reliable foundation.
The hidden challenge: governance
Creating thousands of possible variations is not automatically an advantage.
Without governance, it can create thousands of inconsistent, outdated or unnecessary assets.
To make a digital-twin content system work, companies need processes for:
- Verifying the original 3D model
- Updating it when packaging changes
- Managing regional and language versions
- Controlling access and permissions
- Recording which version is current
- Reviewing AI-generated environments
- Protecting intellectual property
- Checking legal and regulatory compliance
- Measuring whether additional content improves performance
The system must know not only how to create an asset, but also which product version, claim, language and brand rule applies.
This is why the strongest advantage may not come from the AI model itself.
Competitors can access many of the same foundation models and technology platforms. Reproducing a governed library of thousands of accurate products—and connecting it to real marketing workflows—is much harder.
Content becomes a reusable business asset
Nestlé’s wider digital strategy helps explain the investment.
At the time of the announcement, the company said that 72% of its media investment was already digital, that it held more than 340 million first-party data records and that the share of sales generated online had more than doubled over five years. It was moving toward a goal of generating 20% of total sales through online channels.
In that environment, product content is no longer simply a creative output.
It is part of the commercial infrastructure supporting online discovery, advertising, retail-media campaigns and e-commerce conversion.
Every product increasingly needs a persistent digital identity that can be adapted to different contexts.
A digital twin can become the visual foundation for that identity.
What changes for creative teams?
The strongest version of this model does not remove creativity. It changes where creative effort is spent.
Teams can spend less time repeatedly resizing assets, rebuilding packshots or recreating nearly identical regional versions.
More attention can then move toward:
- Campaign strategy
- Storytelling and concepts
- Customer insight
- Cultural relevance
- Experimentation
- Quality control
- Performance measurement
The human role shifts from producing every variation manually to designing, directing and governing the system that produces those variations.
That is a more meaningful change than simply giving every employee access to an image generator.
The bigger lesson for marketing leaders
Nestlé’s use case shows that enterprise AI adoption is moving into a new phase.
The first phase was experimentation: generate an image, draft some copy or test a chatbot.
The next phase is operationalization: connect AI to approved assets, shared standards, business workflows, human oversight and measurable outcomes.
The competitive advantage will not come from producing the largest possible amount of content.
It will come from building a system capable of producing the right content—with the correct product, format, language and brand identity—at the moment it is needed.
Nestlé is not merely trying to create more product images.
It is building a reusable content engine.
That is both an efficiency improvement and a form of infrastructure—and the infrastructure is the part competitors should pay attention to.
Schreiben Sie einen Kommentar