Mondelez International
proprietary genAI platform for ad production
Mondelez invested 40 million dollars in a proprietary genAI platform (with Accenture and Publicis) targeting a 30 to 50% drop in marketing production cost, with social content already live for Chips Ahoy and Milka.
Objective
Reduce the share of the ad budget eaten by creative and production (around 60 percent at a large CPG), produce more variants faster, and spend less on external agencies.
The deployment
Mondelez invested 40 million dollars in a proprietary genAI platform, developed with Accenture and Publicis Groupe. The group targets a 30 to 50 percent drop in the production cost of marketing content, on a creative-production line item that accounts for around 60 percent of a large consumer-goods maker's ad budget. The first uses are already live: social content for Chips Ahoy and Milka, Oreo product pages on Amazon and Walmart. For now the group avoids images of people and focuses on animated product visuals, with human review of everything the AI outputs. The roadmap pushes toward short, ready-to-air TV spots for the 2025-2026 holidays, with the stated ambition of a Super Bowl 2027 format. One cited example: an eight-second Milka animation where waves of chocolate roll over a biscuit, the kind of shot that traditionally costs hundreds of thousands of dollars.
Results Proof C
Investment and objectives announced by Mondelez through its executives, picked up by CNBC and the trade press; the cost reduction is a stated target, not yet an audited result, hence C.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- plateforme Plateforme genAI proprietaire Mondelez Internal ad production platform (40M dollars invested): animated product visuals and videos, with a human review workflow and exclusion of images of people.
- llm Modeles de generation d'images et de video Generative models built into the platform; their names are not published in the sources.
- integrateur Accenture Co-development of the proprietary genAI platform.
- integrateur Publicis Groupe Co-development of the proprietary genAI platform.
How it runs, concretely
For ops teams-
1Brief and framing marketing / brand
Definition of the product, format, and message per market, fed into the proprietary platform.
-
2Variant generation AI / platform (Accenture, Publicis)
Production of animated product visuals and videos (e.g. the Milka animation), in several versions.
-
3Human review Mondelez creative team
Quality and compliance check of each output before release; exclusion of images of people.
-
4Distribution marketing / e-commerce
Go-live on social and e-commerce product pages, with progressive extension to TV.
The fidelity of the product rendering and adherence to brand guidelines. The guardrail held: no images of people, human review of each output before it goes live.
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
- product assets and brand guidelines
- packaging references per market
Org prerequisites
- a team able to run a genAI platform and handle its review
- a make-or-buy trade-off against external agencies
Possible stack
- genAI platform (in-house or via integrator)
- image and video models
- human review workflow
The plan, step by step
- Step 1Pick a high-volume scope (social content and product pages for one brand) and gather product assets, packaging, and guidelines.Deliverable: Brand corpus ready and a backlog of formats to produce.
- Step 2Stand up a POC with an integrator or an internal team on a batch of product visuals, setting the guardrails (no people, systematic human review).Deliverable: First visuals validated by the brand and a written guardrails charter.
- Step 3Industrialize the generation -> review -> publication workflow and measure cost per asset before/after.Deliverable: Tooled pipeline with documented cost per visual.
- Step 4Extend to other brands and markets, then to short video formats.Deliverable: Catalog of multi-market variants in production and a savings assessment.
First step: Target social content and e-commerce product pages, where volume and repetitiveness justify automation.
Sources
- S1 Mondelez invests $40m in generative AI to cut marketing costs Secondary archive pending
- S2 Oreo-maker Mondelez to use new generative AI tool to slash marketing costs Established press 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.