The Architecture of Intelligence: How Hershey’s is Rebuilding the Media Buying Playbook with AI

While the advertising industry has been fixated on the hype surrounding autonomous, "agentic" media buying bots, a quiet, tectonic shift is occurring beneath the surface of brand marketing. The true revolution in performance advertising isn’t just about automated bidding; it is about the fundamental restructuring of how brands measure, model, and forecast the impact of their investments.

Hershey’s, the iconic American confectioner, has emerged as a vanguard in this space. Over the past year, the company has moved beyond mere experimentation, integrating sophisticated AI-driven Media Mix Modeling (MMM) into the core of its business strategy. By treating AI as a structural foundation rather than a flashy add-on, Hershey’s is redefining the relationship between marketing spend and bottom-line growth.

The Shift from Tonnage to Relevance

For decades, the primary metric for media success was "reach"—the sheer volume of eyes on an advertisement. However, in an increasingly fragmented digital landscape, Hershey’s is championing a philosophy of "relevance over reach."

Vinny Rinaldi, the company’s vp of media and marketing technology, argues that optimizing for cheap CPMs (cost per mille) is often a "cheap thrill" that obscures actual business value. "You can gamify any system to go after tonnage for a lower cost," Rinaldi notes. "It doesn’t mean you’re standing out."

This shift has forced Hershey’s to re-evaluate its entire measurement ecosystem. The core question for the brand is no longer "How many people saw this?" but rather "Who was this truly intended for, and did it drive a meaningful change in their behavior?"

Chronology of a Digital Transformation

The transformation at Hershey’s did not happen overnight. It was the result of a deliberate, year-long infrastructure build that prioritized technical foundations over quick-win tools.

• 2018: Hershey’s establishes direct ownership of its ad-tech stack, securing its own contract with The Trade Desk. This set the stage for a future where the brand, not the agency, holds the keys to the data.
• 2021–2023: The company begins a three-year intensive effort to customize its bidding ecosystem using advanced data sets, partnering with platforms like Chalice to integrate custom algorithms across YouTube and Meta.
• 2024: Hershey’s formalizes its relationship with Publicis Groupe while deepening its technical work with AI partners like Mutinex and Tracer.tech.
• Present Day: The company is now utilizing a proprietary system of AI agents capable of modeling campaign performance in weeks, a process that historically took months.

Supporting Data: The Reddit Case Study

The efficacy of this new, agent-driven approach is best exemplified by Hershey’s recent pivot toward Reddit. Through its modernized modeling system, the brand identified that while Reddit might not be a "traditional" mass-reach play, it offered a level of consumer alignment that outweighed traditional cable metrics.

"What we learned with Reddit quickly is that they’re moving more units per dollar spent than any other channel," Rinaldi explains. By analyzing how specific subreddits interacted with brands, Hershey’s was able to craft a strategy that felt native to those communities, leading to superior conversion rates.

This move is part of a broader portfolio diversification: Hershey’s currently allocates approximately 72% of its budget to digital channels and 28% to traditional media. The AI-driven modeling is now being used to inform the current TV upfront season, helping the brand determine exactly where to reallocate dollars if it decides to pull back from legacy cable television.

The Role of Agentic Technology in Planning and Buying

Rinaldi emphasizes that while the industry is obsessed with "buying agents," the real value lies in "planning agents."

"If you only focus on buying, you’re only focused on reach and cost measures," Rinaldi says. "If you’re focused on relevance, you hone in on your communication strategy—what you’re saying to a consumer, how you’re speaking to them, and where you’re speaking to them."

By feeding hyper-focused data into AI models, Hershey’s is testing agents that assist in the complex planning phase—a domain traditionally dominated by human guesswork and manual spreadsheets.

The Question of Ownership: Brand vs. Agency

A critical takeaway from the Hershey’s model is the insistence on brand-owned infrastructure. Rinaldi is clear: because the AI is fueled by the brand’s proprietary sales data, the intellectual property must belong to the brand.

"We own every single technology and data contract," Rinaldi asserts. "The agency just sits in our buying seats in our ecosystem, but they are still our agents-of-record." This "triangular relationship" between the brand, the agency (Publicis), and the platform (The Trade Desk) ensures that the brand remains the architect of its own data strategy.

Implications: The Unsexy Truth About AI

The most significant implication of Hershey’s approach is the sobering reality of implementation. While the media landscape is dominated by the allure of generative AI like ChatGPT and Claude, Rinaldi warns that the infrastructure required to make these tools effective is fundamentally "unsexy."

"It’s like building a home: it can be the most beautiful thing on the outside, but if you forgot to pour the concrete foundation, the first storm is going to blow it over," he says.

For brands looking to replicate this success, the lesson is clear: skip the hype and focus on the plumbing. Building an AI-ready data infrastructure requires years of investment, direct control over tech contracts, and a willingness to integrate sales data into the media planning process.

The Future of Human Critical Thinking

Despite the march of automation, Rinaldi remains a proponent of the human element. He expresses skepticism toward the idea of AI agents buying directly from publishers, noting that there is still a vital role for "human intuitiveness."

"I think critical thinking is one of the most important levers that we have not focused on as an industry in general," Rinaldi argues. By offloading the time-consuming tasks of data aggregation and routine bidding to agents, he hopes to free up agency teams to focus on higher-level strategy.

"If you can replace time spent for both planning and buying through an agentic layer, you’re now elevating the critical thinking capacity of your agency team—who, by the way, know more about the industry than we have time to spend on."

Conclusion: A New Standard for Advertisers

Hershey’s journey signals a turning point for the advertising industry. As brands move away from the "reach at all costs" mentality of the early digital era, they are finding that the most powerful AI applications are not those that simply automate the purchase of ads, but those that provide deep, actionable insights into business performance.

By investing in the foundational architecture of their data, maintaining ownership of their tech stacks, and fostering a collaborative relationship between human intuition and machine-led modeling, Hershey’s has established a roadmap for the future. In this new era, the most successful brands will not be those with the biggest budgets or the most "autonomous" bots, but those with the most coherent, data-driven, and structurally sound approach to the science of marketing.