Mondelez is reshaping their relationship between its global brands and their agencies. Its custom generative-AI platform—built with Publicis and Accenture—is already delivering 30–50% reductions in creative production costs while accelerating content supply across markets and formats. With TV-grade assets planned for the 2026 holiday season and even Super Bowl consideration for 2027, the company is demonstrating what happens when creativity becomes a scalable capability rather than a network of external services. This week we delve into this very interesting case study on how rubber is starting to hit the road in terms of GenAI-enabled workflows.
Focus On: Beyond Cost Reduction and The Velocity Imperative
What does it mean for a brand when creative production moves from a craft-driven workflow to a modular, engineered system?
The £40 million Mondelez investment offers an early blueprint. Rather than displacing creativity, the platform reorganises it—breaking production into components that can be optimised, reused, and tested with far greater precision. Initial deployments, ranging from Chips Ahoy social content in the US to Milka animations in Germany, show that high-quality creative and industrial-grade efficiency can coexist when supported by the right operating model.
The timing matters. With tariff pressures rising and discretionary spending tightening, marketing budgets are being asked to do more with less. Traditional agency structures—optimised for billable hours and campaign cadences—struggle to meet these demands. Mondelez, by contrast, is treating creative production as an internal competence to be developed, governed, and scaled.
The 30–50% cost reduction is juicy, but the real breakthrough is velocity. An 8-second Milka clip with multiple audience-specific backdrops produced at a fraction of the traditional animation cost signals a new economic logic. Creative variation, once expensive, can now be generated and tested in near real time.
Expansion into product-detail pages for Amazon and Walmart further illustrates the strategic intent. These high-volume assets have historically been constrained by economics; Mondelez can now dynamically adapt PDP videos to seasonality, product range, and retail context. The result is a faster, more responsive content engine that strengthens commercial execution.
Equally important is the governance layer. The platform incorporates human review and explicit safeguards around health messaging, overconsumption, and stereotypes. Mondelez has recognised that brand safety at algorithmic scale requires design, not after-the-fact correction.
The Agency Paradox: Partners Become Platform Enablers
WPP’s launch of Open Pro—giving brands the ability to develop ads directly through AI—demonstrates that the agency landscape is shifting. Publicis and Accenture’s role as co-architects of the Mondelez platform reflects a pragmatic repositioning: value is shifting from content production to platform capability, integration, and governance.
This raises important questions. For brands: do you have the organisational maturity to operate creative infrastructure? For agencies: can you evolve from project execution to capability development?
Coca-Cola’s recent experience with AI-generated holiday campaigns—technically polished but emotionally weak—underscores why human ingenuity and oversight remain critical. Technical proficiency alone cannot carry a brand narrative. Mondelez’s measured approach, centred on oversight and governance, acknowledges this.
Your 90-Day Preparation Roadmap
A structured approach helps organisations move from interest to impact:
Phase One: Foundation (Days 0–30)
Target high-volume, low-risk formats such as product-only animations, social variants, and retail media assets. Model the benefits against cycle-time improvements as well as cost-per-asset metrics; the strategic value often lies in speed rather than pure savings.
Phase Two: Capability Building (Days 30–60)
Blend external partnership with internal ownership. A creative holding company can help translate brand systems into machine-readable logic, while an IT integrator can ensure security, governance, and interoperability. Maintain ownership of the underlying capabilities.
Agency platforms such as Open Pro can serve as transitional solutions, offering faster deployment and lower capital requirements. The priority is maintaining architectural flexibility so today’s choices don’t constrain tomorrow’s evolution.
Phase Three: Operational Integration (Days 60–90)
Define the human roles that make an AI-enabled creative organisation function. Roles such as creative operations lead, automation producer, and brand safety reviewer become essential. Rework procurement to reward outcomes rather than hours. If agency partners resist, it signals structural misalignment.
A New Measurement Framework
Traditional creative KPIs remain relevant but incomplete. The new operating model requires metrics that capture the dynamics of AI-enabled production:
Creative Volume Elasticity: Map the relationship between asset variation and performance lift. Early studies suggest diminishing returns beyond 15–20 variants, though this varies by channel.
Speed-to-Market Advantage: Measure total deployment time, not just production time. Competitive advantage increasingly lies in reacting faster than the category.
Cost-Per-Outcome Evolution: Track cost-per-engagement, cost-per-conversion, and ultimately cost-per-incremental-sale. This reframes creative as a revenue engine.
This shift creates winners and losers. Roles focused on mechanical versioning, asset coordination, and routine adaptation will contract. New roles—creative technologists, automation producers, AI-ops managers—will expand. Organisations that succeed will be those that redesign workflows around human–machine collaboration rather than attempting to graft AI onto legacy processes.
Mondelez’s approach is instructive: start narrow, implement guardrails early, maintain human oversight, and scale capabilities progressively. The question for leaders is no longer whether this model will define the industry, but who will move early enough to benefit from its compounding effects.
Every asset created becomes training data. Every variant tested improves the system’s understanding. Every cycle accelerates the next. This is creative production as a learning system.
Follow me
That’s all for this week. To keep up with the latest in generative AI and its relevance to your digital transformation programs, follow me on LinkedIn or subscribe to this newsletter.
Disclaimer: The views and opinions expressed in Chronicles of Change and on my social media accounts are my own and do not necessarily reflect the official policy or position of S&P Global.