How Kraft Heinz measures AI project value

Companies are honing their approach to identifying the right AI use cases and demonstrating the associated business value.

In the case of Kraftheinz, the success of the initiative is measured by a combination of quantitative and qualitative assessments. Pat Nestor, Head of Decision Intelligence Products and Platforms With the food and drink giant.

“On the quantitative side, there's a business case evaluation,” Nestor told CIO Dive. Teams need to be able to connect use cases to broader business goals, such as increasing sales and reducing costs.

Quantitative assessments are combined with qualitative values, such as their effects. Net Promoter Score and recruitment level.

“We measure and track NPS, recruitment and engagement scores to provide a healthy feedback loop and refine and evolve,” Nestor said. “That qualitative quantitative balance is what we really focus on here.”

With a larger budget that focuses on more C-Suite, investors and board-level AI and supports its efforts, technology leaders are under pressure to get the formula right as they choose scalable and beneficial use cases. Track your success after deployment.

AI projects failure rates are rising The companies are working on this year A set of deterrents. Almost everything – 97% – oF According to an Informatica report published in February, F companies are struggling to demonstrate the business value of generated AI.

Successful companies often Credit use cases integrity with broader strategic goals to support project prioritization and take into account available resources and expected timelines.

“One of the most important things for us is, in my advantage, we don't see AI, or in particular generative AI, as a solution of its own, but as a key ingredient that can recharge larger programs,” says Nestor.

Before Kraftheinz chooses a place to place a big AI bet, there is an upfront effort to put pressure on the test use cases. This is also the stage when success criteria are established and mockups are created.

Nestor said the team will first try to answer the next question:

• Where does the company want to go to use cases?
• How can it potentially evolve?
• How is it measured?
• How can I assess my progress along the way?

“That high-value activity funnel is so important that there is a balanced idea and a balanced business case before actually committing material resources,” Nestor said.

However, according to Nestor, a failed experiment or pilot phase is not the end of the world. Learning from failure is part of any innovation process.

“The bottom line means you have to expect and accept some degree of failure,” Nestor said. “We need an innovation funnel, not an innovation tunnel.”

Immerse yourself in AI

Earlier this year, Kraftheinz reorganized its technology governance structure. The company is one global technology umbrella Corrado azzarita, the company's global CIO.

As part of the restructuring, Nestor has shifted from his previous role as Kraftheinz's global product head, leading the newly formed decision intelligence product and platform team.

“Essentially, it's data to AI value chain organizations, so all the facets and sub-teams within it are part of our decision intelligence portfolio,” Nestor said. The team also manages platforms and products within the domain.

Reorg coincides with a broader strategic shift as the company grapples with a six-quarter decline in total revenue. Food Dive reported last month. Kraft Heinz aims to do a lot of levers to turn back the decline in sales. Unlocking efficiency and increasing productivity.

To achieve these goals, Kraft Heinz relied on its strategic technology partners. This includes Microsoft as part of the manufacturing and Snowflake's digital twin efforts, and as the foundation platform for the internal AI engine Khai.

According to Nestor, Khai took about 180 days from the idea and is currently in the hands of around 13,000 team members worldwide. The company wanted to democratize generated AI access without sacrificing security. The AI ​​engine was built internally and built to provide “one-way doors.”

“There's still a wealth of information that Khai is trained and the underlying LLM is being trained, but once team members interact with the tools and upload Kraft Heinz inside information, they won't be able to go out,” says Nestor.

Team members use the platform to analyze and summarise documents and dense industry reports, as well as optimize and access standard operating procedures on factory floors. The engine also helped streamline the deployment of China's sap and automate daily finances. Nestor said he has a platform bookmarked in his browser and uses it every day.

“It's really becoming an integral part of a variety of workflows.” Nestor added that he hopes to see a growth in the number of active users.

The company expects previously referred capabilities Colloquially as Kraftgpt It is fully integrated into Khai. This tool is now called Plantchat and is used throughout the manufacturing ecosystem to aid in strategic decision-making.

Kraftheinz has several other AI pursuits The supply chainincluding among them Pickle Process.

“We believe that each of these of Kraftheinz employees has both quantitative and qualitative value,” Nestl said.