Our mission? Help PepsiCo employees understand how this analytical powerhouse could transform their daily decision-making without making them feel like they were being replaced by a very smart spreadsheet.
PepView 2.0 was elegant in its approach: sophisticated analytics that made humans smarter, not obsolete. Employees could dive into market data, run scenarios, and make informed choices about product development. The human stayed in control while technology served as the ultimate wingman.
At the recent NIQ c360 conference, Bobby Watts, SVP at Ahold Delhaize, perfectly captured what’s keeping CPG executives up at night. He highlighted three critical trends: AI’s operational and personalization impact, the evolution beyond traditional ROAS measurement, and the push for more dynamic, personalized creative. What struck me was how these trends intersect at exactly the point where agentic AI becomes both incredibly powerful and potentially problematic.
Here’s where things get interesting (and slightly terrifying). Today’s agentic AI doesn’t just crunch numbers and offer suggestions. It makes decisions, takes actions, and runs with ideas—all while you’re potentially grabbing lunch or sitting in yet another meeting about meetings.
CPG companies are already testing systems that autonomously optimize pricing, generate marketing campaigns, and suggest new product concepts. According to recent surveys, 71% of CPG leaders have welcomed AI into at least one business function—a jump from 42% in 2023. Gartner’s calling agentic AI the top tech trend for 2025, which means we’re about to see AI systems become as common as that one colleague who always microwaves fish in the break room.
This is exactly what Bobby was getting at when he talked about AI’s impact on personalization and “the elimination of traditional audience building.” Agentic AI can process market data at superhuman speed, iterate through thousands of product concepts, and optimize everything from formulations to packaging designs while considering supply chain constraints and competitive positioning. It’s like having a product development team that never sleeps, never gets cranky, and never argues about the office temperature.
But here’s the catch Bobby didn’t mention: without proper human oversight, we might end up with Salt and Vinegar Donuts on grocery shelves.
Picture this: market research data automatically triggering creative campaigns. Consumer insights flowing directly into packaging design tools. Analytics platforms chatting with creative software like old friends at a reunion.
We’re probably 18-24 months away from this reality becoming commercially viable. NIQ data could seamlessly connect with Adobe Firefly, turning market insights into visual concepts faster than you can say “brand guidelines.”
The problem? Current AI creative output is about as predictable as Florida weather. Sometimes you get brilliant innovation that makes you wonder if the machines are secretly attending design school. Other times, you get results so bizarre you question whether the AI has been binge-watching surrealist art documentaries.
While AI can generate thousands of packaging variations, it can’t reliably tell the difference between “breakthrough innovation” and “career-ending mistake.” That distinction still requires something AI doesn’t possess: taste, cultural understanding, and the ability to recognize when an idea should never see the light of day.
This connects directly to Bobby’s point about creative needing to be “more dynamic, personalized” while still being effective. AI can certainly make it more dynamic and personalized, but effective? That’s where human judgment becomes non-negotiable.
Bobby also hit on something crucial that most people are overlooking: measurement. He pointed out that “not every campaign is created equal, and nor should they be measured the exact same.” This becomes exponentially more complex when AI is making autonomous decisions about product development, pricing, and creative execution.
If your AI system decides to launch a limited-edition flavor based on social media sentiment analysis, how do you measure success? Traditional ROAS doesn’t capture the brand equity impact of a product that either delights customers or becomes a meme for all the wrong reasons. We need measurement frameworks that can evaluate AI decisions in real-time, not just after the quarterly results come in.
This is where agencies like HDco become more essential, not less. The future belongs to companies that master AI orchestration, not just AI deployment. Think of us as the conductors of a very talented but occasionally rebellious orchestra.
Our evolving role includes:
Creative Operations Management: We’re developing systematic approaches to make AI tools reliable enough for brand work. Because consistency matters when your brand is on the line.
AI Training and Prompt Architecture: Teaching AI systems to understand your brand voice, compliance requirements, and market nuances. It’s like brand training, but for robots.
Quality Control Systems: Creating approval workflows that blend AI speed with human judgment. Speed without wisdom is just expensive chaos.
Human-AI Workflow Design: Just like we did with PepView 2.0, we help teams adopt new technologies without losing their humanity (or their minds).
Our experience with PepView 2.0 taught us that successful technology adoption isn’t about the shiny new tool—it’s about change management, workflow integration, and ensuring humans feel empowered rather than obsolete.
These lessons are directly applicable to agentic AI implementation. The most successful CPG companies will:
Agentic AI will transform product development in ways we’re only beginning to understand. It will accelerate innovation, optimize processes, and unlock possibilities we haven’t even dreamed of yet. But it will also require sophisticated human oversight to ensure innovation serves real consumer needs rather than generating technical curiosities.
The companies that nail this balance—leveraging AI’s analytical superpowers while maintaining human judgment and brand wisdom—will dominate their markets. Those that don’t might find themselves in boardrooms explaining why they greenlit Pickle-Flavored Ice Cream or Caffeinated Cereal.
The future doesn’t belong to companies that replace humans with AI. It belongs to those who create the most effective human-AI partnerships. And that’s where agencies experienced in translating complex technology into human-friendly workflows become indispensable.
At HDco, we’ve been helping brands navigate technological transformation since before “AI” was the answer to every business challenge. We understand that the best technology amplifies human creativity rather than replacing it.
Ready to explore how agentic AI could transform your product development while keeping the human insight that drives real innovation? Let’s build your roadmap for AI-enhanced—not AI-replaced—product development. Because the future should be built by humans and machines working together, not by machines working alone.
