General Mills' Digital Roadmap: Agentic Infrastructure, Digital Twins, AI Personas
General Mills has been on a digital transformation runway since 2020, implementing end-to-end solutions that now layer agentic and cloud-based capabilities onto a data framework built for customization.
The company is seeing significant benefits as a result of these upgrades, according to statements from General Mills' supply chain and digital innovation leaders, shared during a recent investors conference.
Building a Digital Transformation Roadmap
General Mills' tech strategy has revolved around three pillars: technology, process and people. On the technology front, it moved operations to the cloud, built a connected data foundation and upgraded its core SAP systems to S4, said Jaime Montemayor, chief digital and technology officer.
Since 2021, all of the company's tech has been running on the cloud. And the data foundation has enabled General Mills to accelerate its investments in data, analytics and AI.
"In process, there was a strong focus on data governance and ethical systems, and you will see how that investment that we made since the beginning of 2020 is paying off today because our data is perhaps as clean as it needs to be for us to accelerate AI investments," he said.
Paul Gallagher, chief supply chain officer, added that this governance structure hasn't been just around the master data that goes into its SAP system, but also operational and transactional data, resulting in a data accuracy rate of about 97%.
"We've got one version of the truth. It's great to have the data, but the second part is that you have to use the data … let the data make the decision," said Gallagher, stating that this approach has so far yielded $300 million of savings over the last three years.
Data-Enabled Modernization in Action
Over the last several years, General Mills has launched a variety of use cases built on its cloud- and data-enabled infrastructure.
Project Elf: An end-to-end logistics flow that uses generative and agentic AI to talk system-to-system, improving communication with one of General Mills' biggest retail partners and optimizing truck utilization.
Gallagher said it has reduced over 15,000 tons of carbon to date because they have fewer trucks on the road. "What used to take us 18 hours to go through those orders … to get truckloads … is now taking us less than 30 minutes."
Growth Labs: A process that uses visualization tools to create and iterate hundreds of visual prototypes so General Mills can better understand consumer needs before manufacturing a physical prototype.
Digital Personas: The company created digital buyer personas with help from its sales teams to gain feedback on innovations, combining real-world learning with the digital insights and AI-moderated research platforms that can conduct hundreds of consumer interviews overnight using chatbots.
Breeding Programs: General Mills has implemented advanced machine learning into its oat breeding program, which sources ingredients for brands such as Cheerios, to predict traits before seeds are even planted.
Regenerative Agriculture: The company uses satellite imagery to improve measurement and modeling for operations that impact greenhouse gas emissions.
"We can now harvest the volumes of technical data that are available to us — not in months, but in weeks and days and hours — to help foster and forward our thinking on how we apply said technology to our application," said chief innovation technology and quality officer Lanette Shaffer Werner.
Predictive Models: General Mills created algorithms that allow it to retrieve insights early, reducing time spent in the inventory system by over 50%. The company expects to increase this to 75%.
"We are using predictive models and real-time data to optimize things like serial line speeds," said Werner. "We are combining a physics-based twin model with our control systems and some in-line sensors to create a foundation that we know we can redeploy across other platforms, such as pet."
Product Lifecycle Management: Upgrades in this space have helped streamline product formulation changes and improve data accuracy and traceability.
It's an effort that helps get more of General Mills' data connected end-to-end, said Werner. "It's resulted in about a 25% reduction in touches when we go in and have to make a change, and also time savings in our conversion of specs and specifications that come from my world and how we translate those into bill of materials at our plants."
The company plans to expand investments such as these, especially through its largest technical center yet. General Mills recently broke ground on its James Ford Bell building, which will increase its available pilot plant capacity by nearly 25%.
Montemayor said the next frontier for General Mills is expanding its tech foundations by adding real agentic AI architectures — rather than just core AI (machine learning or generative capabilities).
"For us, that's the dream — to move from a world where we enable the business through descriptive analytics to a world where we enable the business to prescriptive analytics," said Montemayor.
