How AI is transforming the food and beverage industry

We’ve all been there: juggling family schedules, meetings, deadlines, and a flood of texts, only to get home from the grocery store and realize we forgot a key item on the shopping list. 

Soon, AI might remember to not only supply everything on our list, but also purchase items we may not have even known we wanted, automating the process based on personal tastes, dietary preferences and more.  

According to the 2025 CPG survey by PwC, 40% of consumers say they expect to use AI for comparison shopping by 2030, and an estimated one third of those consumers expect all of the purchasing decisions to be automated. For millennials, the audience considered most likely to use this technology, 62% of those surveyed say they expect to order more online and 46% say they’ll use automated purchasing apps, reordering through smart devices with predictive technology. 

By communicating with a chatbot (with optional voice commands) powered by GenAI, an agent can assist with a personally-tailored shopping list based on an individual’s new product requests, past purchases, and might even anticipate the next round and collect product samples based on video-watching habits like cooking demos. Through various online interactions, it can recognize spending habits, personal tastes, and even budget thresholds, delivering an autonomously-conceived, value-driven haul of goods that is purchased using comparison pricing algorithms with product reviews. No need to wait for sales or deal with coupon codes, the AI shopper will instantly get the best deal. 

In the next few years, smart homes are projected to be what smart phones were by the 2010s — a technology increasingly embraced as the norm by the mainstream. Home appliances, not just laptops, phones and tablets, but the refrigerator, will know when to replenish supplies like milk and coffee, as well as keep track of dietary restrictions and other considerations (such as penchants for sustainably-made products and reduced microplastics) when reordering and building lists. PwC says, “In five years, the phrase ‘grocery run’ may sound as dated as ‘dial-up.’”

One of the standout features of this technology is its knack for predicting what people might buy next, drawing on their online habits. It can even sync with devices like smartwatches to build shopping lists shaped around wellness goals. For instance, if a user has a robust workout routine, the AI agent might look for breakfast options that are more protein-rich, or calming teas for bedtime. 

However, according to the survey, there is a high demand for safeguards, with Gen Z respondents the most concerned about controlling the settings. They want to approve all purchasing decisions, and 100% money-back guarantees are a crucial sticking point. They also favor receiving alerts for shopping receipts and insist on automation only for pre-approved item categories. 

In addition, survey participants want the AI assistant to provide rationale for unexpected purchases and expect a feature that would allow them to shut off the technology at any time, avoiding any “I’m sorry, Dave. I’m afraid I can’t do that. You will eat two dozen sorghum protein bars and you will like them” scenarios. (A reference to the famous line from rogue AI robot HAL in the 1968 sci-fi classic film “2001: A Space Odyssey”).  

Incidentally, 32% of those surveyed say they would never allow the AI assistant full financial access. 

The goal in five years is to deliver the most advanced, user-friendly, result-driven technology to ease, possibly eliminate, the usual hassles of shopping. According to PwC: “Consumers aren’t looking for features, they’re looking for solutions and outcomes. Products are no longer one-dimensional. They’re contextual.”

Major food companies adopting artificial intelligence programs for product development and quality assurance say the technology platforms are helping them gain an advantage in a competitive and fast-paced industry while also staying two steps ahead of potential market disruptions.

Kellanova and ingredients giant Ingredion are among those companies leveraging the technology in its supply chain to refine recipe formulations and make sure products are up to date in accordance with regulations.

The two companies detailed their use of AI along with artificial intelligence provider Tastewise during a virtual State of Innovation in Food Manufacturing event hosted by Food Dive and sister publication Manufacturing Dive on April 9.

AI throughout the entire supply chain

David Lestage, chief R&D officer at Kellanova, said the snacking giant behind Pringles and Rice Krispies Treats is using the technology to tackle operational, regulatory and consumer challenges. Kellanova in 2024 listed artificial intelligence as one of the five top tech priorities it planned to double down on, along with digital twins and data analytics.

Among other things, Kellanova uses AI to identify which ingredients can be substituted if particular raw materials become unavailable. The platforms the company uses, Lestage said, would be able to identify different grades or variants of rice or corn.

“You’d be able to switch things out much more efficiently and have less downtime, less labor burden as well, and accelerate decisions to make those changes,” Lestage said.

Kellanova and Ingredion also are increasingly using AI to keep up with regulatory compliance, particularly as President Donald Trump’s trade war creates new uncertainties. To identify the latest regulatory developments from across the world related to sectors like packaging and carbon emissions, Kellanova uses a platform called RegAsk. The tech allows the company’s workers not to get too bogged down with trying to untangle every country’s unique regulations, according to Lestage, which allows employees to focus on other tasks.

At Ingredion, Chad Davis, vice president of global supply chain, said the ingredients provider is using AI technology to predict which ingredients would be most affected in the face of a supply chain crisis, like tariffs or unexpected speedbumps, such as a ship getting stuck in the Suez Canal.

Ingredion created a “digital twin” of its supply chain network, Davis said, with its manufacturing and distribution sites, supplies of materials along with their costs and consumer trends data from around the world. The technology allows the company to see how certain events like tariffs would ripple throughout the supply chain, keeping Ingredion prepared if potential challenges come to fruition.

“We have all of that modeled in a global simulation. So as an interruption occurs, like a tariff, we update tables in our simulation, and it gives us the next best solution,” Davis said.

 

Predicting the next big consumer trend

Beyond the supply chain, companies also are relying on AI to keep up with ever-changing consumer trends to tap into the next big flavor or product launch.

Kellanova’s food scientists use Tastewise to expand the company’s creative abilities, like drafting packaging, claims and recommended recipes. It also uses Microsoft’s AI tools in platforms like Edge and Copilot to mine insights for product development, Lestage said.

There are nuances and upsides the Cheez-It maker has found in each platform it has tried. “The tools now that allow you to get consumer feedback… accelerate your product development timeline tremendously, by months,” Lestage said.

AI company Tastewise analyzes billions of data points, such as social media interactions, home recipes and restaurant menus to forecast which trends are bubbling up to the surface in the food industry.

The platform includes a chatbot similar to ChatGPT, which allows users to pose queries to specific questions. This process allows companies to quickly predict what consumers will demand rather than just “static research trends,” Tastewise COO Tal Tochner said.

“Companies that are actually winning are the ones that are able to personalize and really understand their consumers,” Tochner said. “They’re asking super hyper-local insights or audience-tailored questions in order to tailor their marketing strategies and win the shelf.”

The Claussen pickle brand traces its roots back more than 150 years ago to an Illinois farmer named Claus F. Claussen who had a truckload of cucumbers he couldn’t sell. Instead of tossing them, he turned them into pickles.

Today, making pickles is a big business. Brand owner Kraft Heinz processes approximately 60 million cucumbers annually to make roughly 42 million jars of Claussen, the country’s top-selling refrigerated pickle brand. But making the crunchy spear isn’t always easy.

The cucumbers that turn into Claussen pickles move from vine to brine in 10 days or fewer, giving Kraft Heinz little room for error. It’s paramount that Kraft Heinz knows what the cucumbers coming into the Claussen plant in Illinois look like so it can prepare — varying circumferences, lengths and bends can wreak havoc on planning and require changes to the production line where the spears are processed.

In recent years, Kraft Heinz has incorporated artificial intelligence across many facets of its supply chain to increase efficiencies and wring out costs. Cucumbers and its Claussen pickles brand have been among the biggest beneficiaries of the once-futuristic technology.

Bill Durbin, the head of North America logistics and planning at Kraft Heinz, sat down with Food Dive to discuss the impact of artificial intelligence on the company’s supply chain and Claussen, one of its top-selling consumer brands in its respective category.

This interview has been edited for brevity and clarity.

FOOD DIVE: How did Kraft Heinz get to the point where it could implement AI in its supply chain?

BILL DURBIN: When we talk about the overall company transformation, ... the supply chain played a very critical role within that, making sure that we're continuing to evolve to drive efficiencies, to then be able to reinvest back into marketing and into our businesses to be able to continue to grow revenues and improve our performance.

For the supply chain transformation itself, I think as we started, it was really about getting brilliant at the basics and improving our processes and identifying best practices, then rolling that out across our 60 manufacturing facilities, warehouses and distribution centers. And so a very manual effort to get us there, but we saw a lot of benefits by identifying the best way to do something and then standardizing that across and continuing to raise that bar.

But over the last few years, we really kicked off this digital journey where we've been able to accelerate that transformation. By leveraging better tools and improving the visibility of exceptions that happen within the network, and then leveraging things like machine learning, we've been able to get people out of those manual transactions so that they can then help drive further optimization of the supply chain, versus having to be in that day to day transaction.

When did Kraft Heinz first start using AI for cucumbers and how did the idea come about?

DURBIN: We started this about a year ago. Our cucumber supply chain is a very short one. It's 10 days from the field to when it comes into the jar. During certain times of the year, we source cucumbers from fields closer to the manufacturing facility, and during the winter, we source from warmer regions which leads to longer transit time. In a lot of our processes, we can buffer with inventory if the goods are received and they're not exactly the specification.

With pickles, the circumference matters, the size matters, the length matters, the bend of the cucumber, all of those things, depending on what they are, we will operate differently within the site, as well as quality. If the quality does not meet our specifications, we can't run it.

We started by bringing in batches of cucumbers, taking pictures of those batches and the quality team was getting the feedback on which ones were to specification, like validating, sizing them. We were training the machine to do that same task.

Similar to that prior example, over time, we trained and then saw different types of defects, different types of sizes, what those things meant. And then [AI] learned from that, beginning to identify these variations, and the operator would validate those things. So now it's able to identify those things on its own.

By being able to do that, we can, by having that level of certainty on what that product is, we can address that right away so we know where to send it within the factory, or if it's an issue, we can get that real-time feedback to the suppliers to be able to address.

Today, we're doing this at the factory where the cucumber becomes a pickle in the jar that goes in the case, and we're starting to explore how we introduce this technology earlier in the supply chain so it's where the cucumber goes on the truck. There we get more time or more responsiveness on being able to work with the supplier or to identify the sizes coming in so that we can schedule the factory correctly for those different sizes.

How is this benefitting Kraft Heinz and Claussen?

DURBIN: On pickles, specifically, we've seen, since we put this in place, we've seen a 12% increase in efficiency from that. By being able to make this process and identify these things, we've been able to make sure that the pickles are getting routed to the right place to give us the best efficiency possible, and also to give that feedback to the suppliers.

We are getting better quality from the start. That's allowed us to get more cases out of that factory as we've improved that efficiency, which means better service for our customers.

Are you using a similar process for any of the commodities Kraft Heinz uses? 

DURBIN: We're starting to look at those opportunities as well, so we haven't put them in place. For tomatoes, it's the size, it's the color. And so we're looking at tomatoes, potatoes for Ore-Ida French fries on how we can now scale this across those other operations. This was really about building this. There was, to be honest, almost a greater need on the cucumber because of how truly short that shelf life is.

Oreo cookie maker Mondelēz International created a new generative artificial intelligence tool to help it personalize its advertising for consumers while boosting engagement for many of its top brands. 

The snacking giant behind Chips Ahoy!, Ritz and Perfect Bar started working on the generative AI tool known as AIDA (AI + Data) more than two years ago, and so far has spent upwards of $40 million on the technology. AIDA enables Mondelēz to create marketing content faster and at a lower cost, often giving it the opportunity to personalize the material for specific consumer groups.

But while the upfront cost is high now, Mondelēz expects the tool could cut the cost of creating marketing content by up to 50%. It could save the company even more in the long term if it is implemented into other parts of the food manufacturer’s business.

Mondelēz, which launched the nascent platform in July, is still learning where and how to best use the technology across its sprawling worldwide snacking portfolio. The tool is still being tailored to understand the intricacies that come with each brand and how to remain responsible in advertising by avoiding the promotion of unhealthy behaviors such as overindulgence.

Jennifer Mennes, vice president, global head of digital marketing and strategy with Mondelēz, and Tina Vaswani, the company’s vice president of digital enablement and consumer data, recently sat down with Food Dive to discuss AIDA and the role of artificial intelligence in food marketing.

This interview has been edited for brevity and clarity.

 

FOOD DIVE: How long has Mondelēz been working on AIDA and why was it something the company believed would be useful for its business?

Mennes: We've been extremely thoughtful of how we tackle this because the entry investment is quite high. So in order to make this a priority for the marketing organization and Mondelēz overall, we need to make sure that we were very thoughtful of the different types of features that we needed to build, that we're going to drive the most value back as quick as possible, which also made us consider what brands we should pilot first. 

But ultimately the decision is that the volume of content we have to produce to really fulfill the end-to-end marketing ecosystem to drive our ambition around personalization, to drive high-level engagement and conversion really requires an entirely different level of volume of content. Doing it the traditional way today, it was not going to be attainable. So we had to find automated solutions, AI being one of them, to deliver against the content volume ambition, to really make sure that we are able to engage with consumers at the fidelity and the volume that we need to improve our business.

It's an enabler. It's not a net-new strategy. It just allows us to do more faster and better.

Vaswani: Part of the process was also re-envisioning how we do the work currently, and then seeing where introducing AI would really provide an uplift or an assist to drive efficiency. That's super important, because we've all learned this from our own experience, that just applying AI on top doesn't always give you the best results. 

So even as we're looking at features, we're being very thoughtful and mindful of really assessing, is this really a value add, or is this actually more of a time drain for the engineers to try to develop the technology where it may not be ready yet.

Are there applications you have found where AIDA has been particularly effective? Similarly, are there areas where it needs some tweaking or places where something is better left to humans?

Mennes: It's always difficult not to get ahead of your skis on what is the expectation of the output versus what can the maturity of the technology actually deliver. We learn so much every day, like what we can produce for a biscuit or a cookie is very different than the outputs that we can get from chocolate.

We focus our attention on areas where we can drive at more speed and scale. But it is a lot of experimenting, experimenting on a very large scale, but every iteration, every prompt is an experiment to see how far we can push the system.

Like on Oreo, we only trained it on the black and white sandwich cookie, the original, our teams can get golden Oreos out of the system without having to train [the AI system.] They're pushing the system to see how far we can – honestly, I call it, break it until we can make it – how far can they push the system to drive as much value out as we can. 

Are there any challenges working with AIDA because it focuses on a specific food product? 

Mennes: A lot of our CPG peers, the product they show, it's a bottle. They're not showing the inside products. We show the product in order to deliver on the taste appeal and the impulse. It's about the product. It's less about the packaging. 

So to make sure that you're adhering to the fidelity of the product, to keep the taste appeal high, the expectation of these models is much higher than if you're just showing a shampoo bottle. It doesn't feel that different, but in the actual training of the data, in the fidelity that you need, it's a night and day difference.

Vaswani: The one thing that we’re learning in real time is that AI has a lot of great promise, but we have a big responsibility to stay true to our brand and to the quality of the imagery that our consumers are accustomed to. And so this is where we're realizing that AI has a bit more to go in terms of getting it perfectly the way we [want it.]

Are there areas where you have had to train AI after it suggested something that doesn’t fit with Mondelēz or a specific brand?

Mennes: Responsible AI is not just about trademark and copyright. It also needs to adhere to our principles, we don’t want to show overindulgence, so I can't have outputs of 18 cookies or 15 pieces of chocolate.

We make sure that there are also brand rules associated with that, so that we are keeping within their own framework. We haven't tackled some of our more regulatory brands, like Halls [cough drops] or Belvita [biscuits], where we have claims. If we did, we would build rules into the platform that you can't say this, you can only say it this way, so that when the systems get prompted, they already know those rules to make sure that the outputs meet those requirements.

Nothing goes into the market without a legal overview. We're not going around the advertising approval process. So once the asset is ready to go into the market, it has to go through the manual legal process that we have today. So at no point is this automated and syndicated out into the market. It just allows us to get there quicker, by making sure that we're not inadvertently putting in things like being a little bit more overindulgent, or just using language that will get flagged in a legal review anyway based on our current processes.

Are there plans to bring AIDA to any other parts of Mondelēz’s business?

Vaswani: If we look at AIDA and the infrastructure underneath, it's definitely scalable. What we're working on is defining what are our next two, three, four big value cases, and then understanding where is it that this would fit under the suite of AIDA and where there is a new platform that needs to be built. We're being very thoughtful and pragmatic. We've built a foundation that is expandable, but how we expand it depends on what are the next big value pieces that will determine how we expand the current infrastructure.

Most enterprises struggle to decide which AI use cases to pursue and feel stuck in pilot phases. The right strategy can help technology leaders light the way. 

PepsiCo is beginning to hit its stride regarding project prioritization as pressure mounts to show results, according to Athina Kanioura, the company’s EVP and chief strategy and transformation officer. 

“There isn’t one rule that fits every company," Kanioura said. "But what we’ve done, which has proven successful so far, is give people room to play while having a very focused agenda on what we call enterprisewide generative AI capabilities that are critical for the company’s future.”

Kanioura said this dual approach helps to promote innovation and employee engagement without overextending resources or getting bogged down by a billion pilots. 

“We will not stifle innovation, but we also need to move toward getting the big ROI on those investments in the places that matter most,” Kanioura said. 

The food and beverage company established an internal sandbox where employees can experiment and test out tools. The platform, called PepGenX, was integrated with Amazon Bedrock to increase flexibility and capability for application development as part of PepsiCo’s multiyear agreement with AWS.

Most businesses cast a wide net when generative AI began gaining traction, holding hackathons and soliciting ideas for implementation. More recently the focus has shifted to sifting through use cases to find what best aligns with broader business goals — and it hasn’t been easy. More than half of IT and business leaders say it's challenging to choose the right use cases based on metrics like costs, business impact and the ability to execute, according to a Snowflake survey. 

As companies abandon most of their AI initiatives and pare back pilots, technology leaders face pressure to move projects along faster and transition them into production.

Establishing key priorities has helped PepsiCo stay focused.

“We have four or five big bets, and that has helped direct investments and align teams,” said Kanioura, such as using AI to help rethink innovation and workforce management. The company embedded generative AI into the product lifecycle management process to manage information about formulas, recipes and deadlines for different launches. 

“Think of a universe of structured and unstructured data,” Kanioura said. “We have now managed to cut the cycle of that whole thing from lots of months to very few months.”

PepsiCo has also used the technology to connect disparate sources, systems and data so employees have easier access to information, from insurance policy details to how to service a laptop.

“This information became much closer to the employee without us having to redesign everything from scratch,” Kanioura said. “The full launch in North America [came] with great results, and we’re continuing the rollout to the rest of the world.”

Kanioura keeps a close eye on the “big bets” and how they track against assigned KPIs. But the internal sandbox is helping the company reach its broader goals, too. 

“It helps people familiarize themselves with the technology,” Kanioura said. “It’s not just a platform to play, you get to understand the value and then you become an ambassador.” 

Moving forward with agents

As enterprises enter the second half of the year, many are eyeing the efficiency gains touted by agentic AI providers. PepsiCo is part of the party, directing resources into AI agents after yearslong investments in digital transformation and subsequent progress on cloud and data goals. 

“People underestimate the importance of investing in the foundation, especially the data foundation, as a prerequisite,” Kanioura said.

Some companies are finding that out the hard way. 

Fewer than 2 in 5 executives have deployed generative AI tools at scale, and just 13% believe the technology has brought significant value, pointing to lagging data readiness as a key culprit, according to an Accenture report published in March. Data woes can erode trust, delay projects and increase costs. 

PepsiCo’s digital transformation journey has relied heavily on a cloud-first approach. Company executives have said the efforts put the data strategy in a better position. 

“If we hadn’t done all those changes and the moves, it would now be impossible for us to do agentic,” Kanioura said. 

While still in the early stages, PepsiCo is exploring how agents can improve the employee and customer experiences. Partnerships are playing a big role in the process. 

“Our internal engineering team cannot build everything, and it doesn’t make sense to,” Kanioura said. “I believe in the right balance between buy and build. You buy what’s a commodity, and you build what is a truly niche competitive advantage.” 

The company has a stringent vendor-vetting process, often courting multiple potential providers before deciding. PepsiCo takes into account a vendor’s roadmap, understanding of industry context, capabilities and more. 

“I need to see skin in the game,” Kanioura said. 

The ability to work with vendors who will allow Pepsi to customize solutions is another big pull for the company.

“Nothing is 100% customized for Pepsi,” Kanioura said. “I don’t believe in extreme outsourcing, and the company always needs to be responsible for its own fate. Having a strong technology backbone of internal software engineers, enterprise architects, data engineers, AI and ML engineers, infrastructure and ops pros is super critical for the success of our organization.”

The holidays are coming, which means a flood of seasonal ad campaigns from major consumer brands and retailers. Leading the way, as it has for nearly a century, is Coca-Cola, a brand whose advertising is inextricably linked to Christmas and the portrayal of Santa Claus.

But for the second year in a row, Coca-Cola is looking to the leading edge of technology to update long-standing holiday advertising traditions. The brand this week released a “refreshed and optimized version” of “Holidays Are Coming,” an ad developed with generative artificial intelligence (AI) that debuted last year (and was itself a remake of a 1995 spot).

While the new ad is again facing backlash, Coca-Cola remains committed to its AI-fueled approach to the holidays, especially since the ad “scored off the charts” with consumers, according to Islam ElDessouky, global vice president for creative strategy and content at Coca-Cola.

“Sure, there is noise and there are people who talk and criticize, but this is one of our top-tested ads in history, period,” ElDessouky said. “The masses, the audiences, do not necessarily look behind the technology. They just look at the story that they’re receiving, and then they respond to it.”

Along with generating high scores in key metrics like association and conversion to transaction, the use of AI also allowed Coca-Cola to utilize a “timeless and timely” framework that balances both foundational brand values and a desire to be experimental and innovative. That risk-taking approach was encouraged by advances in generative AI technology, like OpenAI’s GPT-5, that did not exist last year.

“We’re going to keep going at it, to be honest with you, because it’s giving us the measurements, the metrics and the business results, and, at the same time, we’re learning how to do things differently,” ElDessouky said of AI. “If we do not push ourselves and stretch our comfort zones, people are going to just move without us, and we would love them to move with us.”

Balancing brand needs

Along with the AI-generated “Holidays Are Coming” ad, Coca-Cola’s larger “Refresh Your Holidays” campaign includes a separate, more traditional 30-second TV spot that balances three imperatives for the brand: centering the product, focusing on the heroes of the holidays and connecting to past holiday efforts. 

“A Holiday Memory” does all three, with a narrative about a mother decorating for the season who recalls past celebrations and rewards herself with a classic Coke. The inclusion of a snow globe is a nod to a digital experience from last year’s campaign that turned an AI-assisted “conversation” with Santa into a personalized snow globe asset for social media. 

“We wanted to show that [snow globe] because we’re a brand that lives in every holiday season — things are continuous. We do not necessarily leave ideas behind or anything. It’s all part of our big bank of assets,” ElDessouky explained.

“A Holiday Memory” will run in North America, Latin America and Asia South-Pacific markets — a global approach that speaks to how Coca-Cola is working to personalize and localize campaigns at scale, for both its flagship and other brands in its portfolio.

That approach requires deep listening into insights from different markets and partnerships across human insights, connection and media teams. Research revealed that while multiple cohorts connected with Coca-Cola and Christmas, they had different needs and demands that all had to be weaved into creative. The brand then validates its creative execution before launch and analyzes sentiment once the ads are in market.

“Is this really translating into conversion, either in transaction or transaction and association? We would rather have the latter, but even if we get transaction or association only, it’s still a step in the right direction,” ElDessouky said.

Marketing at scale

Beyond the TV spots, “Refresh Your Holidays” will run across online video, digital, out-of-home, social, in store and on pack channels. Developed by WPP Open X, led by VML and supported by EssenceMediacom, Ogilvy and Burson, the multichannel effort demonstrates the “three Cs” that the brand uses as a framework for campaigns: culture, community and commerce.

For the holiday campaign, culture includes the spots and the out-of-home creative that is created for mass audiences, while commerce is how the campaign is brought to life in retail channels. Community, or where audiences are engaged, includes everything from CRM programs, creator collaborations and experiential activations. The latter includes a tour of the brand’s Christmas trucks in November and December, and is especially critical for reaching young audiences, ElDessouky said.

“The truck tour is actually a nice [thing] that only Coke can do,” the executive said. “A lot of brands have an asset, like the truck, that is synonymous [with the brand]. If you do not capitalize on it… it’s a crime, because if you have an asset, you need to push it.”

Coca-Cola’s holiday assets, from the trucks to Santa Claus and polar bears, remain central to its marketing strategy, especially in the age of generative AI. The company sees the technology as a way to uncover new insights and avenues of engagement that it couldn’t have found otherwise.

“Maybe we land on something that people love so much it becomes our own Labubu,” ElDessouky said, alluding to the viral plushies. “You might land on something like that, and then it becomes an asset of yours, just like we landed on Santa Claus and the truck. Without trying and pushing the limits, we’re not going to add more to the brand.”