The F&B jobs AI is targeting, but is it really that dire?

AI in food and beverage, a summary:

• AI is cutting product development timelines dramatically by modelling millions of ingredient combinations before lab testing
• Automation is expanding beyond production lines into complex tasks, putting pressure on traditional roles
• More than half of industry leaders say AI is already enabling headcount reductions
• Companies are using AI to optimise pricing, reformulation and supply chains in ways not previously possible
• The biggest shift is not full job loss but a rapid redesign of roles towards oversight, data and decision-making

Kraft Heinz had a problem. It was the early 2020s and plant-based foods were surging as more shoppers sought reimagined versions of meat and dairy – predominantly alternatives.

Kraft Heinz’s issue though was a common one: such foods often tasted terrible. Their dilemma was therefore how to avoid a similar fate with its own upcoming plant-based mac & cheese.

In the end, its answer came from artificial intelligence. Specifically, through a partnership with NotCo, a US-based company now helping some of the world’s largest FMCG companies rethink how food is developed.

NotCo’s attraction is its ability to use AI to combine ingredient and sensory data with customer insight and develop new products much more efficiently than the traditional process. In a world where manufacturers are battling not just the competition but supply chain disruption, extreme price volatility, and shifting regulation, an ability to quickly react can be invaluable.

In the case of Kraft Heinz’s Mac & Cheese, a research and development process that would usually take around two years from concept to launch, was done in 10 months.

“The result is a fundamental shift in how product development works,” says Alisia Heath, NotCo’s vice president of R&D. “No human team can realistically evaluate millions of ingredient combinations while simultaneously factoring in taste, texture, cost, nutrition, manufacturing constraints, and consumer preferences.”

For anyone building a career in food technology, that shift may sound unsettling. But Heath insists NotCo is not out to replace people, just “fundamentally augment” them. “Where R&D teams once spent the majority of their time running physical trials, they now direct AI, evaluating candidates computationally before anything reaches a lab bench.”

NotCo’s approach reflects a broader recalibration across the food industry. While early enthusiasm around AI was for flashy, customer-facing robotics - the likes of Amazon Go’s ‘just walk out’ stores or the multi-billion-dollar robotics company Zume which made pizza entirely by robotics and used algorithms to predict customer demand - it’s fair to say such hype has now fizzled out. Amazon announced the closure of all its AI-powered stores in January while Zume is now defunct after pounding through hundreds of millions of venture capital dollars.

Today’s AI adoption is quieter and far more pervasive. According to a recent report by BSI, roughly a third of food businesses now use AI in daily operations. “AI is now embedded across core food manufacturing operations rather than confined to pilot projects,” says Richard Werran, BSI’s global director for consumer, retail and food.

And the impact is felt across the workforce with UK retailers Morrisons and Ocado both recently citing AI as a factor in recent layoffs, while Nestlé pointed to automation in its plans to cut 16,000 roles. More than half of industry leaders say AI is enabling headcount reductions, according to a BSI survey.

Many of the roles under pressure are in traditional manufacturing jobs which, until recently, had broadly resisted automation. Fresh food handling is a prime example where tasks like assembling sandwiches or sorting produce have long relied on a degree of human dexterity due to the variability of ingredients.

What F&B jobs is AI good at?

This is starting to change, however, as advances in AI-enabled machine vision allow robots to ‘see’ and interpret irregular shapes, enabling them to handle delicate tasks with increasing precision. Automation was once limited to highly standardised production lines but is quickly moving into more delicate and aesthetically-driven foods where consistency is critical. Suppliers are in fact already scaling the technology.

Millitec, for example, a major supplier of automation machines for food production, says it deployed its first AI-machine into one of the UK’s biggest sandwich-making factories back in 2020 and it is now making more than 750,000 sandwiches a day.

But industry leaders argue this does not necessarily mean fewer jobs, just different ones. Repetitive handling roles may decline, but demand is rising for technical oversight, systems management and process optimisation. “The manufacturers gaining the most from AI are those using automation to strengthen workforce capability, not simply reduce headcount,” says Nigel Smith, CEO of TM Robotics.

It is often a similar story away from the factory floor where business are deploying AI to unlock capabilities previously out of reach. Tom Holden is the chief product officer at Mondra, a food intelligence company using AI to help Britain’s biggest supermarkets understand the risk and resilience of each ingredient in their products. “We are all about managing price inflation and reducing the likelihood of a stock out,” Holden explains. “The two killer punches that every food company in the world is most fearful of.”

When geopolitical shocks occur like the Iran war, Mondra can quickly model potential disruptions to ingredient sourcing. It can also support product redesign. In one case, UK retailer Tesco used the system to reformulate a ready-meal lasagne, resulting in a product that was more nutritious, lower in carbon intensity and more profitable, Holden says.

None of this was possible before AI, he claims. “They could manage nutrition,” he says. “But if you wanted to extend out into all the dimensions, they did not have the data. It was not available.”

It is an example of how the AI ‘job-pocalypse’ can sometimes be an oversimplification. In this case, AI is not taking the jobs of sourcing managers but aiding them in a developing area with a complexity of data they were previously unable to truly consider.

“It’s a really, really complicated, multi-decisional framework now,” says Holden. “We are beginning to show that using the AI can help you through that.”

AI’s ability to crunch immense data volumes is of course one of its fundamental pillars and in many cases, that is what is bringing most value to food companies. One area consequently gaining significant traction is product forecasting, explains Paddy Winters, a food and drink consultant at Baringa, which while a long-established process is still often a partially manual process even in some of the biggest food companies, he explains.

Now though, it is quickly becoming more sophisticated as the system takes on dozens more data sources and therefore is able to give much greater accuracy. The reduced need for human intervention is also drawing in new users, smaller firms for example, who until now could never afford to invest in such things.

Even some bigger firms are finally coming to the party. “There’s all these businesses out there that supply own-label products but have not really invested in this technology because the business case hasn’t stacked up,” says Winters.

“Labour’s been cheap enough that you can just still do it with people. But because the cost of labour has gone up, it’s changed the economic model.”

This is a crucial point. Because for all the promise and panic associated with AI, there are many in the industry who argue that both sides of the debate are framing the issue incorrectly. They argue this should not be a question of which jobs will AI take, but closer to Winters’ point – which problems urgently need a solution which other means have so far failed to locate.

Will AI take F&B jobs?

As the AI commentator Berend Booms says: “Once you answer that honestly, the role AI should play – and the role people continue to play alongside it – becomes considerably clearer.”

For Booms, the part of the food manufacturing most consistently underserved by the AI conversation is maintenance and uptime. While quality control and supply chain optimisation get a lot of attention, the equipment on the factory floor is where unplanned downtime can quietly cost manufacturers many millions.

AI could analyse failures and machine conditions, he says, allowing maintenance teams to shift from reactive to proactive – “not by replacing their judgment, but by ensuring they have the right information to act on before a line goes down”.

“Even a modest 5% to 10% reduction in mean time to repair represents significant monetary value, quite apart from the productivity benefits,” he says.

There is a clear contradiction amid the rollout of AI across the food industry. While external advisors and commentators consistently argue AI should not replace humans but rather ‘augment’ them, there is already a clear impact on roles from the factory floor up to head office.

Some of it seems inevitable. A food technologist whose role is to determine nutritional information, or a finance analyst running data for a performance review seem at particular risk from AI’s still burgeoning capabilities.

It is now arguably starting to move beyond that though, into realms where many think a human-touch is always be essential.

Jonny Bingham is the co-founder of Bingham & Jones, a consultancy using culinary expertise to help food companies develop new products. He says that while there are still many firms keen to employ people with a palate, who understand food and know how to cook as part of their product development, it is already a diminishing pool.

“If big businesses can get that for cheap then that will always be their default position. Especially when business is difficult.”

Bingham is no luddite, but he offers up a passionate defence for why the humans must stick around: “You go to restaurants or watch TV, and everything is very sexy. Food has become almost porn for the masses. You only need to look on Insta or Tik Tok and it’s full of people doing great stuff with food. But when you reduce it to AI trying to create flavours, tastes, feelings, it doesn’t work.”

He concludes: “AI can give you insight and it can give you a starting place. But to make food really sexy again, it needs a human touch.”