How PepsiCo’s AI Strategy is Dominating Consumer Goods
PepsiCo is deploying AI across its entire value chain in a bid to enhance its market dominance in the consumer packaged goods sector.
The company’s strategy, led by Athina Kanioura, Executive Vice President and Chief Strategy and Transformation Officer, focuses on what she describes as “four or five big bets” rather than scattered pilot programmes.
This approach concentrates resources on enterprise-wide generative AI capabilities that could fundamentally reshape how the company operates.
“People underestimate the importance of investing in the foundation, especially the data foundation, as a prerequisite,” Athina says.
“If we hadn’t done all those changes and the moves, it would now be impossible for us to do agentic [AI].”
The strategy is a departure from the typical corporate approach to AI adoption, which often involves numerous small-scale experiments that fail to deliver measurable business value.
Instead, PepsiCo has committed to major technological partnerships and comprehensive integration across its operations.
CEO Ramon Laguarta acknowledges the company’s data infrastructure was inadequate as recently as 2019.
By 2024, he says, foundational work had enabled a “much more aggressive” data and AI strategy.
Amazon and Salesforce powering PepsiCo’s AI backbone
The company has established partnerships with Amazon Web Services (AWS) and Salesforce.
These alliances form the technical foundation for PepsiCo’s AI applications.
PepsiCo’s internal platform, PepGenX, operates as a sandbox where employees can experiment with Gen AI tools.
The platform integrates with Amazon Bedrock, a service that provides access to foundation models – large AI systems trained on vast datasets to perform various tasks.
Athina describes the platform as giving people “room to play” within a controlled framework.
The system aims to build AI literacy across PepsiCo’s workforce while serving as an educational tool for employees to understand the technology’s practical applications.
The partnership with Salesforce also focuses on deploying Agentforce, an AI agent platform that automates business processes.
This system will be integrated across PepsiCo’s global sales operations, which serve 5 million retail locations worldwide.
The Agentforce implementation synthesises real-time data from warehouses, retailer point-of-sale systems, and consumer behaviour trends.
Sales representatives receive AI-generated insights that analyse variables including regional demand, competitor pricing, local events and weather patterns to suggest localised promotions and stock orders.
Manufacturing precision through machine learning
In manufacturing facilities, PepsiCo uses AI to maintain product quality and optimise operations.
At Frito-Lay plants, which produce snack foods, AI models scan products on production lines and make real-time adjustments to temperature, shape and consistency.
The technology has enabled some facilities to operate for a full year without unexpected equipment breakdowns.
Meanwhile, predictive maintenance algorithms analyse equipment data to shift maintenance teams from reactive repairs to planned servicing.
AI-powered vision systems assess potato characteristics including weight and peel percentage, generating savings of approximately £240,000 (US$307,200) per production line.
In smart factories, AI-enabled sensors and automation have reduced energy consumption by 20% and water usage by 30%.
The company’s procurement operations also benefit from AI integration.
The Salesforce partnership extends Agentforce capabilities to purchasing, where the system creates supplier profiles by synthesising data from multiple sources.
Real-time spend anomaly reporting identifies issues including unauthorised purchases without manual audits.
Research and development acceleration
PepsiCo has compressed new product development cycles from six to nine months to six weeks using Gen AI integrated into its product lifecycle management process.
The PepGenX platform manages structured and unstructured data related to formulas, recipes and launch deadlines.
AI models analyse consumer feedback, market trends and ingredient interactions to accelerate food science innovation.
The technology evaluates ingredient combinations to reformulate existing products, such as reducing salt and fat content in Lay’s crisps whilst maintaining taste and quality.
The company has used these capabilities to launch new brands targeting specific trends, including Off The Eaten Path seaweed snacks and Propel fitness water.
This rapid innovation capability allows PepsiCo to respond to emerging consumer preferences and establish market positions before competitors.
Consumer engagement through personalisation
The company deploys AI for direct consumer engagement through personalised experiences.
The “Lay’s Messi Messages” campaign used AI to map footballer Lionel Messi’s facial features from five minutes of recorded footage, generating over 4 million unique personalised video messages in 10 languages.
Meanwhile, Gatorade, PepsiCo’s sports drink brand, is developing an AI hydration coach named Anna.
The virtual assistant, trained on data from the Gatorade Sports Science Institute, provides personalised hydration advice to consumers. The pilot programme is scheduled for late 2024 or early 2025.
Furthermore, the company has developed a “Smart Can” – a connected device with digital screens, motion sensors and sound technology designed as an interactive portal for personalised experiences.
On university campuses, autonomous “Snackbots” allow students to order snacks for delivery via mobile applications.
Each personalised interaction generates data about consumer preferences and behaviours.
This information feeds into PepsiCo’s unified data cloud to refine demand forecasting, inform product development and personalise future marketing campaigns.
“Much more aggressive,” Ramon says of the company’s current data and AI strategy compared to previous years.
