Kraft Heinz
RAG creative generation engine on a proprietary brand corpus
Kraft Heinz built TasteMaker, a RAG creative generation engine on Google Vertex AI (with Apply Digital), which takes creative design from weeks to hours (about 8x) for a portfolio of more than 100 brands.
Key points
- RAG creative generation engine on a proprietary brand corpus.
- Built on Google Vertex AI with Apply Digital as integrator.
- Creative design timelines from weeks to hours, about 8x faster.
- Deployed across more than 100 brands, evidence level C confirmed.
Objective
Produce on-brand creative content on demand for a portfolio of more than 100 brands, without the delays and costs of standard agency cycles, while keeping control of the brands' intellectual property.
The deployment
Kraft Heinz built TasteMaker, a RAG-type creative generation engine built on Google Vertex AI with the integrator Apply Digital. The system indexes codified brand assets (guidelines, visuals, product references) in an internal corpus served through BigQuery, then generates images, motion, and video on demand while staying within each brand's rules. The instance is proprietary, which protects the brands' IP. Kraft Heinz manages more than 100 brands and reports that TasteMaker takes creative design timelines from weeks to hours, an acceleration of about 8x. The deployment started on four use cases before being extended after the first results. The roadmap adds fine-tuning of image generation on proprietary data, video, and concept testing on synthetic users.
Results Proof C
Several concordant sources naming Kraft Heinz and the 8x factor (Forrester analyst, Google Cloud customer story, integrator Apply Digital, specialized press). No standalone figure in financial results, so C.
How it works
Documented architectureThe stack in detail
- plateforme TasteMaker (moteur RAG in-house) Kraft Heinz's proprietary instance that generates images, motion, and video within each brand's rules, protecting the IP
- plateforme Google Vertex AI cloud AI foundation on which TasteMaker is built
- llm Gemini, Imagen, Veo Google generative models: text and reasoning (Gemini), image (Imagen), video (Veo)
- infra BigQuery serves the codified brand asset corpus (guidelines, visuals, product references) to the RAG engine
- integrateur Apply Digital design and development of the platform with the Kraft Heinz teams
How it runs, concretely
For ops teams-
1Brand asset indexing data team / Apply Digital
Codification of the guidelines, visuals, and product references into an internal corpus served through BigQuery.
-
2Creative request marketing
The brand team describes the need (format, market, angle) to the platform.
-
3Multimodal generation AI
TasteMaker produces images, motion, and video within the brand's rules through Vertex AI (Gemini, Imagen, Veo).
-
4Brand control and distribution marketing
Validation by the brand team, then placement in digital media, on e-commerce pages, and on social.
The corpus of codified brand assets (guidelines, visuals, products) served through RAG. If the corpus is incomplete or poorly indexed, generation drifts from the guidelines and loses brand consistency.
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
- corpus of codified and indexable brand assets
- guidelines and rules per brand and per market
- warehouse-type data infrastructure (BigQuery or equivalent)
Org prerequisites
- quality governance for the generated content
- private AI instance to protect the brands' IP
- marketing team trained in creative briefing on a genAI tool
Possible stack
- RAG engine on multimodal generative models
- cloud platform (Vertex AI or equivalent)
- integrator for the assembly (like Apply Digital)
The plan, step by step
- Step 1Codify the assets of a high-creative-volume pilot brand (guidelines, visuals, references)Deliverable: Indexed and queryable brand corpus
- Step 2Set up the RAG engine on the generative models in a private instanceDeliverable: Platform generating assets compliant with the guidelines
- Step 3Validate on 3-4 concrete use cases and measure timelines before and afterDeliverable: Assets produced + design timeline comparison
- Step 4Set up the brief-generation-brand-control workflow and distributeDeliverable: Process in production across media, e-commerce, and social
- Step 5Extend to the brand portfolio with quality governanceDeliverable: Corpus per brand + common validation rules
First step: Codify the assets of one high-creative-volume brand and measure design timelines before and after.
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 Kraft Heinz Is Using GenAI To Re-Imagine The Future Of Creative Content Secondary archive pending
- S2 Kraft Heinz case study Interested party archive pending
- S3 Kraft Heinz leverages AI to transform creative condiment content Secondary archive pending
- S4 Kraft Heinz Tastemaker Interested party 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.