PUBLISHED DATE: 2026-04-25 04:00:37
Our Ambition: Unlocking Empathy at Scale Through AI-Driven Content Generation and Personalization
The imperative for change
Publicis partnered with a global retail leader faced an “impossible” problem of achieving empathy at scale with relevant, topical campaigns supported by personalized assets. The traditional content creation process required substantial manual input, with the time from campaign inception to in-market assets often extending up to a year and yielding a limited number of assets. Additionally, the brand struggled to consistently refresh its digital commerce content year-over-year, which frequently resulted in stale imagery or text.
Scale and impact – near-term ambition:
The objective was to use AI to accelerate asset deployment while scaling personalization across brand and digital commerce channels. Workflow automation reduced time and resource effort, freeing creative teams to focus on strategy and innovation.
The Future business value opportunity
Although still in its initial launch phases, the platform has proven highly successful with a 98% active user rate and the generation of over 3,500 assets (including images, text, and marketing concepts).
Cost Efficiency:
The solution drove an 8% cost reduction per asset compared to the international snack brand’s historical baseline.
Exponential Volume Growth:
The volume of deployed assets skyrocketed by 200% versus standard business-as-usual metrics.
Responsible Scaling:
The customized compliance engine successfully achieved a 78% compliance rate with responsible AI guidelines for image generation.
The transformative solution visualized
- Advanced Image Generation Pipeline: The solution utilized Google Vertex AI APIs, prominently featuring Gemini Large Image Models like Imagen 3 and Imagen 4. We innovatively implemented Gemini to construct an image generation pipeline that adjusts for culturally relevant outputs and integrates brand guidelines, substantially reducing the need for prompt engineering expertise from the user.
- Custom Post-Generation Editing: To give end users maximum control, we enabled custom editable layered PSD generation by blending the Vertex AI Segmentation API with open-source libraries, allowing for low-level editing after the AI generation.
- Enhanced Compliance Protocols: We developed responsible AI compliance checks that exceeded default Google API protocols, allowing for brand-specific safety and compliance on multimodal data.
- Robust Cloud Architecture: The underlying infrastructure leveraged a comprehensive suite of tools including Vertex AI workbenches, Vector Store, Model Registry, CloudSQL, Firestore, BigQuery, GCS, GKE, Document AI, and Google Artifact Registry.
Sales: Elizabeth Papasakelariou
Delivery & POC: Yogita Sood
Alliance BD: Joseph Lamoureux
Vertical:
Consumer Products
Country:
United States
Customer Profile:
Mondelēz International, Inc. is an American multinational snack food company headquartered in Chicago, United States. Mondelēz is a global leader in the snacks industry, with a portfolio of iconic brands spanning biscuits, chocolate, gum, candy, and baked snacks, sold in approximately 160 countries worldwide.