How General Mills found business value in generative AI

Generative AI has emerged as a seemingly accessible technology for businesses.

Executives quickly realized the potential to improve efficiency and upgrade the customer experience by introducing chatbots, document summarization tools, and digital assistants to everything from call centers to engineering teams.

But in recent years, scaling this technology at the enterprise level has proven seemingly difficult, as AI bias, model illusions, and deep-seated concerns about data privacy have hindered widespread adoption.

“Deploying generative AI comes with a degree of risk for companies that want to use it.” Jaime Montemayor, Chief Digital and Technology Officer, General Millshe said at one point. Wednesday CIO Dive and CFO Dive Live Virtual Events.

Still, General Mills was well positioned to be an early adopter and deployed internal generation capabilities. called AI chatbot mills chat in February.

When the food industry giant began laying the digital foundation for its AI technology, it wasn’t focused on large-scale language models. According to , the company started putting the infrastructure in place to support machine learning several years before OpenAI delivered its first ChatGPT iteration in November 2022. Montemayorhas been selected to lead General Mills i’s digital transformation.February 2020.

As part of its preparation, the company moved its data to the cloud to improve its analytics capabilities. is affiliated with google cloud 3 years ago. General Mills has also had success expanding its ML capabilities.

“We’ve built a lot of these capabilities across the company.” Montemayor Said. “We deploy them in every part of our organization, including marketing, sales and supply chain, and through them we create business value.”

cross-functional value

From cloud and data modernization to AI adoption, executive buy-in for digital transformation efforts helps align IT efforts with business goals. Building relationships between the different departments of the company helps maintain the project once it begins.

“It takes a village for any company to succeed with AI.” Montemayor Said. “One of the first things we did was, a few weeks later, Chat GPT The idea was to form a cross-functional subgroup of our senior leadership team. ”

The management team included heads of human resources, legal, finance, and supply chain. general mills.

Financing is especially helpful as companies move from pilot to full implementation and costs start to mount. In addition to approving funding, finance executives can help IT identify metrics for return on technology investments.

“It’s the finance team that helps keep score.” Montemayor Said. With support from the finance team, the company identifies where to accelerate or reduce investments and new features.

of mills chat Assistant is primarily a writing and summarizing tool. was built using Google’s PaLM2 modeled and distributed to Approximately 900 users earlier this year. This tool is currently in the hands of: 20,000 General Mills employees; According to Montemayor.

The company’s generative AI victory was a result not only of investment in technology but also improvements in its people.

“When we started this journey, we knew we were a traditional company and needed to invest specifically in data science.” Montemayor Said. “Over the past four years, slowly but surely, we have been building the talent base to enable our AI program.”

One of the keys is Montemayor We measure technology capabilities against business priorities. He also emphasized the importance of pragmatism when it comes to emerging technologies.

“I spend a lot of time collaborating with colleagues to experiment and continue to grow the pipeline of work in the core AI machine learning and generative AI areas, and to find a balance between what is possible and what is achievable. I’m sure it’s achievable.”