How are the food giants using AI?

Artificial intelligence is one of the major stories of the decade. It has already radically altered many areas of life, from art and literature to medicine and computer science. Food is no exception.

While ChatGPT was only released in November 2022, many major food companies have already fully embraced AI, using it to make gains in efficiency, R&D, and more.

Nestlé boosts product creation with AINestlé uses AI in many areas across its business. AI is already “embedded” across its value chain, explains a Nestlé spokesperson.

For example, Nestlé’s AI-powered concept generator combines insights on consumers, signals from the market and information about whether or not it can actually be made to develop a product, which radically boosts innovation.

It has already produced 1,300 products, the spokesperson explains, and at least 30 are currently in the pipeline. This has cut development time from three months to three weeks, and may soon be scaled up.

The next step is recipe optimisation, which AI is also doing for Nestlé. When creating recipes, Nestlé uses AI to cut ingredients costs, ensure compliance to global standards, and align with nutritional and environmental targets.

Furthermore, Nestlé is using digital twins, virtual models of objects, systems or processes connected to their real-world counterparts through a two-way flow of information, to adjust or localise marketing content digitally. This will match the product to certain situations, such as seasonal campaigns or channel-specific formats.

“This means that new creative content can be generated using AI, without having to constantly reshoot from scratch,” explains the spokesperson.

Finally, Nestlé is also optimising supply chain efficiency through AI. “In procurement, we are applying AI analysis to the contract base of our suppliers, currently covering over 80% of total spend. Many suppliers cover multiple markets, and AI has enabled us to review hundreds of thousands of contracts for inconsistencies between the global terms of the contract and local application in markets.”

Unilever harnesses AI for digital NPDUK major Unilever is using AI in a range of ways. New product developing is one such example. In-silico testing (testing via computer simulation) allows it to “test millions of recipe combinations in seconds which helps us design products to suit (regional) taste preferences,” explains a Unilever spokesperson.

This means that a lot of the major R&D for products can be done virtually.

“Our modelling capability enables us to do much of the design work for core products now in-silico, from shelf life to microbiological stability, to creating the right texture and taste, as well as predicting how the product will behave on the factory lines.”

Products developed using this method include Hellmann’s plant-based mayonnaise and Hellmann’s real mayonnaise squeeze bottle.

For the plant-based mayonnaise, the company has “developed advanced models that can predict taste, texture and stability of our products. Using our in-silico models, we can replace the animal-based emulsifier (egg) with a plant-based alternative without the need of multiple recipe testing and traditional product development trials.​”

Like Nestlé, Unilever is using AI to build digital twins of its products for marketing purposes.

All the major FMCGs are using AI (Hiroshi Watanabe/Getty Images)Unilever’s digital twins are ‘pixel-perfect’ product images, using technology such as Real-Time 3D, Nvidia Omniverse and OpenUSD, the company creates 3D digital replicas which aim to create an entirely accurate representation of its product.

These product images are then reintegrated back into creative workflows. The digital twins allow the company to save both time and costs in marketing. The company says it’s already seen 55% cost savings and 65% time savings.

Finally, the company is using AI to reduce costs in its factory in Poznan, Poland, saving 10% on utilities (including water and energy) and 20% on cleaning times.

Mondelēz cooks up recipes with AIMondelēz International is using AI to digitally produce product formulations. This has led to the creation of 70 SKUs.

The company’s AI product development (AIPD) tool assists human creators in the development of products, meaning that they are produced between two and five times faster than traditional methods. These are not, the company stresses, AI recipes, and all include a human touch.

Mondelēz is also using its generative AI platform to optimise its marketing capabilities, allowing a faster creation of personalised image, text and video.

The multinational snacking company is partnering with global services company Accenture to do this.

The platform enables the company to collect and process real-time data, which generative AI will then turn into insights that will be accessible by decision-makers across the company.

Such data will allow Mondelēz to aim towards higher degrees of personalisation in its marketing.

Kraft Heinz uses AI to find defective cucumbersKraft Heinz uses AI in a number of ways: keeping an eye on its supply chain via a ‘control tower’; using digital twin technology; and mapping information about ingredients via its proprietary ‘computer vision system.’

This system analyses the quality and specifications of ingredients it receives from suppliers. From a truckload of thousands of cucumbers, it is able to detect a single defective one.

Kraft Heinz is also using Tastewise’s TasteGPT tool to accelerate R&D. Tastewise enables it to keep in touch with current food trends.

Danone bolsters supply chain efficiency using AIDairy major Danone last year partnered with Microsoft to improve its supply chain efficiency.

AI is used to integrate procurement data across several platforms, which generates more complete and enhanced vendor performance, contracts, and spending insights.