Unilever
genAI creative production from product digital twins
Unilever industrialized genAI creative production from product digital twins, reaching about 55% savings and 65% shorter turnaround on beauty, with video completion rate and click-through rate doubled, across 18 markets.
Objective
Move away from advertising production designed for television first and feed the digital channels with product visuals, faster and cheaper, without a systematic reshoot.
The deployment
Unilever is building a digital twin of each product: a single 3D source that holds the variants, the packaging, and the legal text in every language. From that source, teams generate and adapt visuals for paid social, programmatic display, and e-commerce listings without starting from a photo shoot. The Beauty AI Studio, built with The Brandtech Group, serves brands such as Dove Intensive Repair, TRESemme Lamellar Shine, and Vaseline Gluta Hya, and runs across 18 markets. The system does not produce images of people: it works the product, the background, and the staging. Unilever puts the whole at production cost halved and production twice as fast overall, with, on the beauty scope, gains on the order of 55 percent savings and 65 percent shorter turnaround.
Results Proof C
Figures stated by the brand (Chief Growth and Marketing Officer) and echoed by several established marketing press titles; no line isolable in the financial results, hence C rather than A.
How it works
Documented architectureThe stack in detail
- plateforme Nvidia Omniverse Platform to create the 3D digital twins of the products (variants, packaging, per-language text).
- outil OpenUSD 3D description framework used as the format for the digital twins.
- integrateur Beauty AI Studio (The Brandtech Group) genAI production line for beauty, built with Brandtech, in service across 18 markets (Dove, TRESemme, Vaseline).
- llm Modeles de generation d'images Generation of scenes, backgrounds, and variations from the digital twin, without images of people; exact models not publicly named.
How it runs, concretely
For ops teams-
1Create the digital twin data / 3D team + Nvidia (tech)
3D modeling of the product in Omniverse/OpenUSD, with its variants, packagings, and per-language labels.
-
2Generate the visuals Beauty AI Studio (Brandtech) / AI
From the twin, generate the scenes, backgrounds, and variations for paid social, display, and e-commerce, without a reshoot.
-
3Adapt by market local brand teams
Localize the text, formats, and messages for the 18 markets, starting from the same source.
-
4Deliver and measure marketing / e-commerce
Publish on the paid and e-commerce channels, track video completion rate and click-through rate.
The fidelity of the digital twin to the real product (packaging, shade, per-market legal text). If the twin drifts from the product sold, all the generated content becomes false and unusable.
How your customers perceive this type of use
Sourced studiesUn ecart net separe les annonceurs des consommateurs : 77% des annonceurs voient l'IA positivement contre 38% des consommateurs (Yahoo/Publicis, 2024). Les mesures implicites confirment le rejet declare : en EEG, les pubs generees par IA produisent une activation memorielle plus faible que les pubs traditionnelles et sont decrites comme agacantes, ennuyeuses et confuses (NIQ, 2024). La disclosure a un effet ambivalent : elle augmente fortement la confiance quand elle est remarquee (Yahoo/Publicis), mais 27% des jeunes consommateurs disent faire moins confiance a une entreprise dont la pub est creee par IA (IAB, 2024).
Acceptance conditions
- Une disclosure visible : quand la mention IA est remarquee, la confiance globale envers l'entreprise augmente de 96% (Yahoo/Publicis 2024)
- Une qualite visuelle suffisante : les visuels IA de basse qualite augmentent l'effort cognitif et distraient du message (NIQ 2024)
Red lines
- Le contenu IA non declare puis identifie : 72% des consommateurs disent que l'IA rend l'authenticite difficile a etablir (Yahoo/Publicis 2024) et les marques utilisant des pubs IA sont plus souvent jugees inauthentiques ou non ethiques par les consommateurs que par les dirigeants (IAB 2024)
- Les mannequins et personnes generes par IA : 46% des consommateurs n'en veulent pas dans la publicite, l'inquietude premiere etant les standards de beaute irrealistes (Attest 2025)
Sources: Yahoo / Publicis Media (terrain Ebco) 2024 · IAB (avec Attest) 2024 · NIQ (NielsenIQ) 2024 · Attest 2025
How to replicate
Inference, not sourcedData prerequisites
- usable 3D files or product photos
- packaging and legal text reference by market
- brand guidelines
Org prerequisites
- a creative team able to run a genAI pipeline
- a legal validation process for synthetic visuals
Possible stack
- a 3D engine (Omniverse or equivalent)
- image generation models
- an assembly platform like Brandtech, or an in-house stack
The plan, step by step
- Step 1Choose a range with a high volume of e-commerce variations and gather the sources: 3D files or product photos, packaging, and legal text by market.Deliverable: A complete source dossier for the pilot range.
- Step 2Model the digital twin (3D, variants, languages) and validate it against the real product.Deliverable: A validated digital twin, compliant with the packaging sold.
- Step 3Generate the variations for paid social, display, and e-commerce listings, with brand and legal validation of each synthetic visual.Deliverable: A library of approved visuals by channel and by market.
- Step 4Deploy, then compare cost per asset, turnaround, video completion, and CTR to classic production.Deliverable: A cost/turnaround review and a decision to extend to other ranges.
First step: Choose a range with a high volume of e-commerce variations and model a pilot digital twin of it.
The same pattern, proven in other industries
Adobe
in-house genAI production of marketing content at scale
Atlassian
internal genAI agents for account targeting and marketing content production
BMW
advertising and experiential creation with generative AI
Sources
- S1 How Unilever's AI marketing bets are increasing production efficiency Established press archive pending
- S2 Unilever reinvents product shoots with digital twins and AI Primary archive pending
- S3 Unilever is building a gen AI assembly line for its digital creative Secondary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.