How Kroger's AI Factory Is Accelerating AI Adoption
Advances in AI technology continue to create new opportunities for businesses to innovate and enhance operations. However, successful implementation requires collaboration, alignment and often a shift in our ways of working. As businesses strive to develop advanced AI models and technologies, they must also consider effective change management and deployment strategies
84.51° and Kroger are designing an AI factory to accelerate AI adoption across our enterprise. Our goal was to make AI development faster, safer and easier to adopt across all teams as we research, develop and deploy customer-first science in grocery retail, insights, customer loyalty and media.
Accelerating Adoption
The AI Factory is a centralized hub that provides foundational capabilities, services, accelerators and guidance to support AI development across our business. We found that while teams had no shortage of ideas and energy, many AI projects stalled early in development. Proof of concepts were taking too long, and many teams were unaware of existing models or tools they could utilize to get started. We also needed governance and principles to ensure we were building AI responsibly and safely.
To address these challenges, we developed our AI strategy to include a centralized "front door" – a hub where teams can initiate their AI projects by accessing available models, data (excluding any customer, employee or personal information) and tools. This hub also helps teams define the business value of their ideas from the outset and includes assessments to identify potential issues early, such as scalability challenges, adversarial risks or data gaps. I'd like to share a few key insights we learned along the way.
Balance Is Best
Creating an environment where speed and control coexist is essential for sustained and scalable AI progress. While a fast-but-siloed approach can spark rapid experimentation and innovation, it often leads to reduced knowledge sharing, duplicated effort and barriers to scale.
At Kroger, we realized that to unlock full value, we needed to balance bottom-up innovation with top-down guardrails. This meant empowering teams to move quickly while ensuring they operate within clear governance, tooling, standards and security protocols. These frameworks promote the reuse of proven, approved, and secure services – making it easier to scale what works.
The AI Factory plays a key role in enabling this balance. By offering reusable components, feasibility assessments and a developer portal with use case blueprints and APIs, we dramatically accelerated AI delivery. For example, in many cases, starting new use cases is quicker now compared to weeks before. This demonstrates how the right structure fuels both agility and efficiency. Governance and key responsible AI principles limit duplication of effort and safeguard our data – again, balancing speed and control.
The Business Model Matters
Fully centralized AI business models – where all development and deployment happen through a single team – often become bottlenecks, sacrificing speed and domain relevancy. Meanwhile, fully distributed models give each business unit autonomy but frequently result in duplicated work, inconsistent standards and increased risk.
To solve these challenges, we adopted a hybrid "Hub + Spoke" model to get the best of both worlds. In this approach, AI development and deployment happens within the business "spokes," where teams are closest to the problems and best equipped to drive innovation. Meanwhile, the “hub” provides shared infrastructure, reusable components, base models and governance to ensure consistency, compliance and efficiency.
This model allows each business team to focus on delivering impact in their domain – without needing to reinvent foundational capabilities like setting up or managing GenAI models. The result has been transformative: Initial development timelines were drastically reduced, duplicative work lessened and the company is realizing meaningful efficiencies.
Change Management and Upskilling
Standing up the right architecture is only step one. True transformation happens when people, processes and mindsets evolve alongside the technology.
While our tech stack continues to advance – prioritizing open-source, cloud-agnostic solutions – the real impact comes from how we enable the organization to adopt and scale AI. We paired this foundation with intentional change management and L&D efforts, helping teams across the business understand, trust and use AI effectively.
AI is no longer seen as a "bolt on" initiative but as a strategic capability for driving growth, improving customer experiences and empowering associates to solve the right problems with lasting impact.
About the Author
Kristin Foster is SVP, Head of Data Science/AI at 84.51˚/Kroger Precision Marketing. Foster leads a team of data scientists focused on researching, developing and deploying customer-first science in grocery retail, insights, customer loyalty and media. By collaborating with sales, product and engineering teams, Foster's team creates products and solutions with prescriptive science at the point of decision to drive informed business decisions for The Kroger Co. and CPGs.
