Behind the headlines

Automating the food industry has several challenges, according to Eric Benoist of Natixis.

People are aware of automation, but what consumers are often unaware of is the high level of automation that already goes into getting food onto their plates. Eric Benoist, responsible for tech and data research at banking concern Natixis, discussed his findings on how automation in the food processing industry is progressing and what the next trends are, with Dairy Industries International:


The industry has sometimes attracted an unflattering reputation, far removed from that of a sector where technology is at the core of economic progress. In our view, however, the situation isn’t so dark. Much has been achieved already and many customers remain unaware of the impressive amount of technology that sometimes goes into a meal for their enjoyment and satisfaction

We’ve spoken to a few companies, and they all have the same problem. It’s not that easy to find the right solutions and the right robots, when you work in an industry where the source of materials, shapes, colours and sizes, are not the exact same. In the automotive industry, the same piece goes in 6,000 times per hour, but it’s a bit more work in the food industry to make automation consistent.

There are a couple of technology developments. The first one is around machine vision, which is really making good progress in that field. Machines can properly detect whether it’s in good shape or not, can take materials that are quite fragile and handle soft materials without destroying them. In the dairy industry, there have been a few product developments that are quite interesting to do with packaging, which are working perfectly fine.

It is also quite interesting for agriculture and animal health in general. When looking at milking robots, GEA has a solution that’s amazing, which is its DairyRobot R9500. (Ed. Note: The GEA DairyMilk M6850 cell count classes sensor monitors every udder quarter of every cow at milking. This cell count analysis in each udder quarter can detect subclinical mastitis infections, which can benefit directly from faster treatment periods. With a few clicks, a current picture of animal health can be had, or switched to monitoring milking parameters.) This helps to detect problems such as mastitis, which costs farmers €25 billion per annum, so it’s not a bad thing to detect.

Super cooling

Other systems we have been looking into have to do with the cold chain, which is very important in the global dairy industry. The latest development is a cooling technology that is not as cold as usual cold chain, so the products travel a few degrees above the freezing temperature, and it enables a producer to export dairy very long distances with less energy.

Engineering remotely

Traditionally, engineers intervene onsite, install big machines with a very long testing phase, using three to four engineers for eight hours in a row to make improvements, and in person. What they’re doing now is developing digital twins of machines, testing that machine in the digital world, and then importing it onto the factory floor, where then it works from the beginning with less time in person.

Companies are now using digital training and modelling to show how it works on the factory floor, offering remote engineering assistance and ongoing development and upgrades. This requires a bit of adapting and investment.

The problems always include human resistance, as well as different generations of machines on the factory floor. However, it’s important to train people and make sure they adapt to these new environments.

Perhaps even more in the food industry, there are a lot of small family businesses. They need to bring new materials onward, and to work somewhat differently. There is a lot of the human dimension in automation. The question is, are you in a position to train people? Are people trainable? It’s about adapting to this change.


The dairy industry’s going to have to follow the new consumption trends. This depends on what part of the planet you’re looking at, as emerging markets are different compared to the US or EU. In emerging countries, they’re increasing consumption of dairy products that was missing from their diets.

In India and Pakistan, these two countries are quite autonomous, and consume everything they make. For countries such as China, milk is becoming a big thing – they are consuming a lot more of it, interest growing, and this depends on the middle class getting richer quickly. They haven’t had access to packaged food, processed foods and meats, and we see that in milk, it’s going to require lot of automation to improve productivity and hygiene.

In other parts of the world, consumption patterns are changing, with a focus on healthy products – people are more mindful about animal health, and want more transparent products as well. It’s not just a question of automation, a bit more goes into it – real production innovation will be required.

People used to think milk was milk, and it was good to have a daily glass of milk. Now milk can sometimes be perceived as not that healthy, or not good for fat levels. This is a staggering idea, pushed further by new types of alternative drinks, but actually innovation is coming from that area anyway.

Laboratories are going to be able to remove the “bad” ingredients in milk and physically develop products without touching the taste of milk. There is a lot of engineering and artificial intelligence, to understand what creates taste and mouthfeel, to able to produce natural products, while removing items such as lactose that aren’t desirable. There will be more innovation in that space, with new flavours and products for the industry.

There have been developments, such as a computer that’s able to digitise taste to help replace some ingredients with others, while maintaining the taste. For example, in the case of light cheese, it tastes exactly the same, so these products have been able to mix ingredients and replace the mouthfeel – that’s where the traditional industry is going forward. To be able to make cheese taste the same as natural cheese that tastes great, without all the higher fat – that’s important for the next few years.

The next step is producing milk without an animal, where the laboratory manipulates genes, mixes them with bacteria to make a synthetic whey protein, which can be fermented in a tank and then filtered into a powder to generate some kind of milk. I’m not particularly opposed to trying it, but it doesn’t seem very natural. However, it might answer the question of many consumers who want to have “better” environmental and sustainable options. In this never-ending debate, consumers will want answers.

At the end of the day, there’s no point in fighting about things, we need solutions. These questions have to answer the issues arising, all of which can be accelerated

by artificial intelligence and these kinds of technologies. The last point is the question of traceability, which is very important – today’s consumers want to know where the food is coming from, whether the cow is treated properly, and the natural processes.

The problem with blockchain

Blockchain technologies are one of the possible answers. There is a lot of buzz around it, but I am not convinced it works so well for food. When looking at food traceability, there are number of issues there. For example, the data that you put in blockchain, how do you guarantee the information put in is solid? For small farms in small countries that have few frontline IT systems, the farmers may be under pressure to say they’re treated fairly.

The problem with blockchain is the intersection of the physical and digital worlds. With very long supply chains there are potentially dozens of links in a chain, and you can’t guarantee what the label says. There are also very few ways to guarantee everyone’s acting properly. You have to regulate and enforce it, and it only works well if it’s fully automated. For example, you can apply it to the cold chain, to maintain a constant temperature, but that doesn’t tell you that the end product is what you think it is. Interesting.

Automation won’t replace people. We now have great tools helping us answer a number of problems, but they are just new tools for humans to use.

New equipment is extremely capital intensive, and they are developing new technologies to help farmers produce more. So it’s important to be part of the going in the right direction, with a lot for governments to think about as well. There are massive infrastructure questions – satellites, GPS and so forth. This will lead to more precision farming, but perhaps more expensive farming. It is crucial that leaders start thinking about this long-term.

Related content

Leave a reply

Dairy Industries International