Artificial Intelligence

Machine dreams – the future of artificial intelligence in food and beverage operations | Food Industry Analysis


"Its important the food and drinks sector, especially SMEs, start their digital journey with small steps"

“It’s important the food and drinks sector, especially SMEs, start their digital journey with small steps”

Proponents of artificial intelligence say the technology can help food companies, facing pressure from climate change, feed a growing global population with greater efficiency. Lucy Britner looks at how manufacturers could tap into AI.

Artificial intelligence is not an Arnold Schwarzenegger-shaped robot from the future, sent back in time to destroy mankind.

In simple terms, AI is an umbrella phrase for the skills shown by learning systems that are discerned by us humans to exhibit ‘intelligence’. Technology consultant Capgemini says in today’s world typical AI capabilities include speech, image and video recognition, as well as a host of other skills, including complex analytics and predictions.

The use of AI in food and drinks manufacturing falls under what has widely been dubbed the fourth manufacturing revolution, or Industry 4.0. 

In fact, AI has already infiltrated many parts of food and drinks production. From supply chain management, waste management, predictive maintenance and scheduled ordering to weather predictions, food safety compliance and new product development, machine learning is reshaping businesses.

From a big-picture perspective, creating more efficient ways to grow and process raw materials will help feed a growing and ageing global population – and avoid food waste. At the same time, AI is being used to navigate climate change in order to plant crops in places where they will thrive.

At company level, wider use of AI can create cost savings, better business management and therefore more profits. At the moment, the world of AI in food and beverage production appears to be dominated by innovative start-ups and tech company collaborations, developing niche bits of machine learning to tackle specific challenges.

“AI technology can now be applied to every stage of the food chain and can be key to changing these processes, from initial production and farming through to the final products served,” says Nicola Sewell, marketing manager at AI-enabled food waste tracker Winnow. “Those that best harness this technology potential will become the food and beverage leaders of tomorrow.”

Megatrends

The changing population

According to the United Nations, the world’s population is expected to increase by 2bn people in the next 30 years, from 7.7bn in 2019 to 9.7bn in 2050. In a UN report, released last June, the organisation also confirmed the world’s population is getting older, due to increasing life expectancy and falling fertility levels – and that the number of countries experiencing a reduction in population size is also growing.

“The resulting changes in the size, composition and distribution of the world’s populationhave important consequences for achieving the Sustainable Development Goals (SDGs), the globally agreed targets for improving economic prosperity and social well-being while protecting the environment,” the UN report says.

Sustainability

Global food systems are responsible for between 25-35% of human-caused greenhouse gas emissions, according to the WWF. The organisation says in order to address the climate crisis, we need to “significantly change the ways we use land for agriculture and think differently about what we eat”.

Changes, they say, that need to go hand-in-hand with shifting to renewable energy sources and rethinking transportation systems. “And with a growing, hungry, diverse population, we need to do it quickly,” the WWF warns.

Health and wellness

The number of plantbased food and drinks launches has accelerated in the past two years, as more consumers look for alternatives to meat and dairy, in part for health reasons. Meatfree options have officially gone mainstream, with fast-food giants like KFC and Burger King introducing plant-based offerings.

Meanwhile, documentaries such as The Game Changers, Forks Over Knives and Cowspiracy have helped to spur on consumer interest in veganism for reasons of animal welfare and environmental sustainability.

The health-and-wellness trend has also sparked interest in ingredients, as consumers look to understand more about what they are putting into their bodies. This means producers will need to have that information at the touch of a button, and also be confident it is accurate – especially when it comes to allergens.

Convenience

The convenience trend has given rise to meal kits, more sophisticated home delivery and shopping via social media. Consumers have come to expect to be able to get just about anything, at any time. This means food and beverage companies need to be able to predict and anticipate changing consumer needs in order to create the right product and deliver it in the right way.

The technology associated with convenience has also brought with it a wealth of data around what, when, how and where people eat and drink.

Personalisation

Hand-in-hand with convenience, comes personalisation. Not only do we expect to get whatever we want at the touch of a button, we expect it to be just that – what we want.

Food and drinks producers are being tasked with coming up with more personalised offerings, at the same time as finding new and more personal ways to advertise and market to consumers.

The present and future of AI in food and drinks operations

Jerome Buvat, global head of research and head of the Capgemini Research Institute tells just-food a number of large organisations across the world are using AI for “critical parts of their manufacturing operations”.

“Not only does AI help organisations with critical parts such as quality control and demand planning, it is also being used in functions such as NPD”

He says AI is increasingly helping organisations unlock additional value in functions that have been traditionally human-driven. “For instance, not only does AI help organisations with critical parts such as quality control and demand planning, it is also being used in functions such as new product development.”

Buvat gives the example of Danish brewer Carlsberg, which launched an AI project to develop new beers and enhance the quality of existing beers. “Using this system Carlsberg expects to reduce the new product development process by up to 30%,” Buvat says. 

Over the course of the next five years, Buvat believes AI will play a bigger role in food and beverage organisations, becoming a must-have for manufacturers. “It isn’t hard to imagine, for instance, a mobile AI-powered app that can quickly scan a food item and tell both its constituents and whether it is fit for consumption, just through a single image,” he says.

Tip of the iceberg

For Keith Thornhill, head of food and beverage in the UK for global manufacturing giant Siemens, the possibilities for AI in the food and beverage industries have so far “only touched the tip of the iceberg”. 

AI can be used from farm to fork, he explains. “There are AI solutions that can be used using the Internet of Things (IoT) to predict crop yield,” he explains. “For instance, with Siemens’ MindSphere solutions, our clients can capture data, like weather conditions during the life cycle of a rice crop over a period and analyse which climatic conditions yielded the best result. This will allow farmers to plan their crops accordingly. The variability factor in weather conditions can be learned and the life cycle of crops managed to yield a bumper crop.”

Thornhill believes the next five years will see food manufacturers looking to AI in order to shorten the time to market for their products.

Waste management

At the ‘fork’ end of ‘farm-to-fork’, using AI to cut down on food waste can help to drive sustainability and efficiency. At Winnow, Sewell describes food waste as an “age-old problem” for foodservice operators. She says Winnow data from kitchens around the world suggests 5-15% of food purchased ends up in the bin.

“Kitchens that use Winnow tend to see a 40-70% reduction in food waste within six  to 12 months, driving food cost savings in the range of 2-8%, improving margins whilst doing the right thing,” she adds.

By using a camera that is trained to recognise foods that end up in the bin, Winnow is able to offer precise data accuracy to enable kitchen teams to more easily identify waste patterns.

Outside help

There are also funded schemes to help companies get a leg-up when it comes to embracing new technologies. Last year in the UK for example, food firms were among the businesses to sign up for a government-backed initiative called Made Smarter.

The GBP20m (US$26m) pilot centres around providing advice to SME manufacturers keen to adopt technologies such as AI, virtual reality, IoT and sensors, 3D printing and robotics – but don’t know how to go about it.

Among the companies involved in the scheme is Eggbase, a software specialist working to collate and share data from the egg and poultry industry. Eggbase is building a platform that would allow data to be collected, analysed and, by applying AI to the dataset, create new data products for the industry.

The platform aims to improve bird welfare – monitoring health, preventing disease – as well as improve practices and efficiencies through better decision-making. “Eggbase is a good example of how you take that first step into AI,” explains Kevin Smith, industrial digital technology advisor at Made Smarter. “It’s a project born out of the demands of UK egg producers for quality data and predictive analytics.

Eggbase aims to collate data into a central warehouse in order to – in the future – train models or machine-learning algorithms to achieve more intelligent decision making.”

Smith describes the initiative as “the virtuous circle of AI”, expanding thus: “Collect data, train a machine-learning algorithm, deploy it, get data back, build a better product, get more users and get more data, and it goes round again.”

Continuous learning

Symphony RetailAI, which specialises inAI-enabled decision platforms and insights for retailers and CPG manufacturers, has also worked with food and drinks manufacturers, including meat behemoth Tyson Foods, beer major Anheuser-Busch InBev and UK manufacturer Premier Foods. 

“With AI, learning from past marketing and promotional efforts is continuous”

“Ideally, in a fast-moving market, manufacturers would be aware of any market shifts or problems before they actually happen, and traditional waysof leveraging data do not allow for this,” explains Jonathan Tye-Walker, client director at Symphony RetailAI. “It is simply humanly impossible to manage the massive amount of data a CPG company has. With AI, learning from past marketing and promotional efforts is continuous.

Tye-Walker believes future AI trends fall into a number of silos. However, he says AI technologyand a personalised understanding of a company’s customer base are a common foundation.

“For example, as customers become increasingly environmentally-focused, as veganism increases and as the population shows signs of becoming an ‘ageing population’, it will be paramount that retailers and manufacturers are able to understand changing shopper trends quickly and communicate with their customers at a hyper-personalised level in real-time,” he adds. “Building in-market, in-store activities around a dynamic customer-centric strategy is a must.”

Overall, Made Smarter’s Smith believes the trajectory for AI in the food and beverage sectors will be similar to other manufacturing industries. “As they start to collect and build their data warehouse it gives them the option to do more things in the future,” he says. “As the knowledge of what is possible becomes more widespread through programmes like Made Smarter, we will start seeing more applications.”

In two years’ time, Smith feels certain there will be many more examples of training algorithms to simplify processes, predict maintenance or project market demand.

What can food and drinks companies do now?

Embrace change

The food and beverage manufacturers that will flourish are those set up to embrace the wave of change that new technology will bring to the market, believes Symphony RetailAI’s Tye-Walker.

“Everyone has witnessed the changes to retail in the high street we see today and the impact of growing omnichannel and digital retailers,” he says. “The world is changing and the new decade looks set to bring a wave of AI-driven innovation to retail and to their customers; be that through prescriptive analytics, increasingly effective promotional forecasting and dynamic in-store and online assortment – all underpinned by dynamic, supply-chain demand forecasting utilising multiple data sources to create predictive insights.”

Tye-Walker believes manufacturers that are ready to collaborate with retailers will be the winners in the future.

Start small

It is important that the food and drinks sector, especially the SMEs, start their digital journey with small steps,” says Siemens’ Thornhill. “Firstly, automating their food production units and then incorporating bigger technologies like AI.

“The journey of innovative technologies for any organisation can be daunting but with the right approach it can become a smooth transition.”

Scaling up

Capgemini’s Buvat says food and drinks companies need to ensure their implementation of AI moves beyond proof-of-concept (POC) stage, which remains a sizeable challenge for manufacturers.

“Scaling AI is critical since it gives organisations valuable expertise, which can be re-used and reapplied in future AI projects,” he explains.

Buvat recommends deploying successful POCs to live engineering environments: “A live engineering environment will help the AI system learn and im prove accuracy,” he says. “This, along with integration with existing IT systems, will ensure a smooth transition as the system is made live in more locations.”

For Buvat, investing in a strong foundation of data also goes hand-in-hand with investing in AI systems and talent. “A strong data and talent foundation will allow organisations to maintain momentum when the value of AI has been proven by the first few use cases,” he says. “It also helps in creating repeatable, faster, and easier roll-outs of new AI applications in the future.”

Lastly, in order to enable seamless scaling to the broader manufacturing network, Buvat says organisations should move use cases to an AI platform, accessible through the organisation, to capture the full value of available data and resources. “This also allows for centralised access and portability of these AI applications,” he adds.

Collaborate

In January, Danone announced a partnership with Microsoft to launch AI Factory For AgriFood.

The initiative aims to encourage projects serving regenerative agriculture (soil health, animal welfare, support for farmers), sustainable food, waste minimisation and optimisation of supply chains.

Cécile Cabanis, Danone’s CFO as well as the French group’s executive vice president for information systems, cycles and procurement, says the Activia and Evian owner believes artificial intelligence can contribute to what the company calls the “food revolution” – a necessary shake-up of the global food system to make products healthier and more sustainable. AI, Cabanis says, can play its part “by improving our agricultural systems and our food’s value chains”.

Elsewhere, some universities offer collaborative programme when it comes to boosting technology in business. For example, Carlsberg’s project is a team effort – along with Aarhus University, The Technical University of Denmark and Microsoft.

Prepare for a challenge

It is no surprise big companies are looking to start-ups or collaborators to help kick-start their AI journeys. The arena is still nascent. “Aim high and expect a challenge,” says Jochen Förster, director and professor for yeast fermentation at the Carlsberg Research Laboratory. “The technology is still being developed. It’s very early days for everyone and our main challenges have been that AI is still at an early development stage.”

The wrap: while AI might look intimidating in the movies, its application in food and beverage manufacturing has the potential to reduce waste, drive efficiency and save money. 





READ SOURCE

This website uses cookies. By continuing to use this site, you accept our use of cookies.