You’re reading AI at the Top - Our new series on Sundays, where we share how leading companies and their executives use AI in business.
If you enjoy your soft drinks and chips, there’s no way you have missed a PepsiCo product.
In fact, that’s an understatement. There is a high chance you debate over favourite flavours with your friends.
With over 1.4 billion sales a day, PepsiCo is one of the largest food and beverage companies.
Plus the brand has all the numbers you’d expect from an enterprise. 400,000 retail outlets, 250,000 employees, $196.50 billion market cap, 60 petabytes of data growing at 2x a year, you name it.
In today’s AI at the Top, we will learn about PepsiCo’s AI partnerships, use cases, and tangible results at its massive scale.
What is PepsiCo’s AI play?
Unlike JPMorgan Chase which built most of its servers, cloud, agents, and tools in-house, PepsiCo invested in its tech enterprise partners to strengthen the AI game.
Think AWS, Salesforce, NVIDIA, and even select startups for niche solutions.
The company’s focus is on enterprise-wide value chain integration instead of excelling in a specific domain like its competitors.
Coca-Cola’s AI focus is on marketing and CX. Nestle is into supply chain and R&D.
PepsiCo believes its end-to-end strategy connects the entire value chain with a single, intelligent thread and a common set of objectives.
So… What did PepsiCo implement with its collaborations and strategy?
1/ PepGenX
In partnership with AWS, PepsiCo created its internal Gen AI tool PepGenX.
The company considers it a sandbox that gives its employees the room to play and experiment.
If you work at PepsiCo, you can use PepGenX to choose multiple models, build applications, and generate reports in minutes (enterprise employees never appreciated faster reporting more; such a relief.)
PepGenX also democratizes AI across the company and tests which employees are not reluctant to change, who would later become AI ambassadors within the enterprise.
2/ Agents
PepsiCo collaborated with Salesforce’s Agentforce to build the company’s agentic use cases.
Athina Kanioura, PerpsiCo’s Chief Strategy and Transformation Officer mentioned, “With Agentforce, rich data is enabling better decision-making and efficiency across our organisation, paving the way for a more resilient future-ready enterprise.”
Today, the company uses agents for:
Gathering data from any source to create unified customer profiles
Product stocking optimisation in real-time
Faster, responsive CX
Quick in-store execution
Deeper insights into user behaviour and patterns for targeted marketing campaigns
3/ Speaking of Marketing Campaigns…
PepsiCo’s Gatorade launched a marketing campaign, asking its customers to design their own bottle by describing the design.
Results? 150,000 unique designers were generated with a 165% increase in weekly sales and a 200% lift in loyalty.
This is not a new idea; we know personalisation works and PepsiCo doubled down on it. It’s a proven strategy with brands like Nike and Nutella launching similar campaigns.
While campaigns like Gatorade’s hit headlines, here are the boring use cases of business that move the needle massively
One of PepsiCo’s large use cases is in procurement, to ensure the raw materials’ quality isn’t compromised. Like potatoes to make Lays.
The brand ensures the supply chain is always updated with fast, real-time information. Magesh Bagavathi, PepsiCo’s SVP, Global Head of AI, analytics, and data, mentioned today’s AI is about the persona-centric approach.
“Knowing the workflow of each person in our supply chain allows them to optimise each step efficiently.”Demand forecasting happens at two stages: R&D and store management. In R&D, AI allows the brand to forecast the need for ‘healthy chips.’ In store management, say, the retailers know when to stack up the refrigerators with Gatorade because there’s a heat wave coming.
PepsiCo had its challenges developing a strong, reliable forecasting system in retail stores.
For example, the training data would include a neatly packed refrigerator, but these are images AI has to process in the real world:
Yet PepsiCo processed over 10 million images to improve its accuracy in stocking and retail management.
This is a dynamic process with multiple iterations.
🧠 Enjoying this read? Share the smartest part with your followers. Click to share.
At this point, we are off the honeymoon phase
We know it’s not about the ‘wow factor’ anymore with AI, but what results AI will help us meet.
Before we end this edition, here are some tangible results from PepsiCo’s AI:
🌟 Want to be featured in the next issue? Reach out with your best AI use case and we’ll spotlight it.
If you enjoyed this deep dive, you should definitely read our editions on Walmart, JPMorgan Chase, and more.
Sources:
Case Studies by AI Magazine, Manufacturing Digital, JE Logistics, AI Media House, IFT Org, and Klover.
Magesh Bagavathi’s Interview and AI platform solutions team presentation.
YOUR TURN