How artificial intelligence is helping supermarkets predict their sales

A small grocery store in the countryside of Germany had a bit of a challenge. Being far from the big city, the store mostly relied on local high school students for its workforce. According to Avik Mukhija, one of the founders of Freshflow, a retail tech startup, these kids were eager to work but lacked experience. He’s keeping the exact location of the store under wraps.

Every day, when the young workers ordered fresh fruits and vegetables, they tended to order more than needed because they were worried about running out. This resulted in unnecessary waste because some of the produce would inevitably spoil and go unsold in the store.

But then the store’s bosses started using a machine learning system from Freshflow. This system can predict how much fresh food customers will buy in the next few days and suggest what stock to bring in. With this tool, the amount of produce wasted at the supermarket dropped by almost 30%. Freshflow has 15 employees and has raised €3 million. Their software is being used in several stores owned by two local grocery chains in Germany, and they’re also testing it out in France.

Also Read: AI-generated art explained: What is it and how does it work?

Reducing the food waste

For a while now, managers of fresh produce in supermarkets and small local shops have had to figure out on their own how many apples, strawberries, or potatoes to order, just to keep up with demand.

But demand is always changing. In-store deals, holidays, the state of the stock already on the shelves, and the weather – all these factors can affect sales. The person ordering fresh stock has to make sure they get their numbers right every day.

In the EU, about 60 million tonnes of food waste is produced each year, with a market value of over €1 billion. Most of this waste comes from households, while retailers like supermarkets make up a smaller portion. However, the European Commission still expects retailers to do their part in reducing food waste.

Tracking food’s freshness automatically

A bunch of new startups are saying they’ve got a better way to handle the flow of fruits, veggies, meat, and fish in and out of stores. They’re talking about technology that can cut down on waste and make more money. These companies say that artificial intelligence (AI) and detailed data about how long stock lasts on the shelves could really change things in the food retail world.

Freshflow also keeps an eye on how often product managers listen to the advice from its software – and right now, it’s about 90% of the time, says Mukhija.

The system takes in all the data Freshflow can find. It’s not just the weather, but also sales numbers that are almost up-to-date, or how well certain deals, like buy-one-get-one-free offers, have done before. The software can look at how stores are doing in terms of waste and also show which products are selling the most at any given time.

Mukhija says they didn’t make Freshflow to replace the job of a human fresh produce manager. “This is something that we feel very strongly about,” he says. “We will never replace the fresh produce manager. We will only enhance their abilities.”

He says it’s a job that needs “art” – more than he and co-founder Carmine Paolino knew when they started their company. The fresh produce manager often finds new ways to show off or advertise the fruits and veggies for sale, trying to get the local customers interested.

That changes depending on the store, and Freshflow’s software isn’t made to replace that job – it’s just supposed to make it easier to decide what produce to order, since there are so many things to think about.

Tracking food’s freshness automatically

Other startups are looking at how to keep track of how fresh food is without needing people to do it. “We came up with a gas-sensing technology,” says Max Grell, who started BlakBear, a company in London with 11 people that’s raised £2.3 million in funding, not counting grants and other private money they haven’t talked about.

Grell says that BlakBear’s small sensors, which he says will be really cheap to make a lot of, go inside food that’s been packaged – like bags of fresh fish or meat, which are important sources of protein. The sensors can tell when the air inside those sealed bags changes a little bit as the food gets older.

Grell won’t say exactly what gases the sensor can find, but he says it can find things like ammonia, volatile organic compounds, and carbon dioxide. If the sensors notice more of these gases, which can mean the food is going bad, they can send a message about it using wireless RFID or Bluetooth.

“We’ve spent a long time working on protein because protein’s super expensive,” says Grell. “It can make people sick. It smells really nasty when it’s bad.”

Right now, BlakBear is working with two big stores in the UK, says Grell, but he won’t say which ones. They’re also using their technology in the US. By 2025, Grell wants to have “millions” of sensors in food packages moving around the supply chain.

Also Read: Top five AI trends the banking world could see in 2024

Replacing the use-by date

Right now, the sensors are going into a few packets instead of every single one, so stores can check the quality of all the food they get. But Grell says his company’s big plan is even bigger: “The vision for us is we replace the use-by date.”

That would mean every bag of fresh fish or meat, for example, would have a sensor in it that could tell the customer – if they scan it with a smartphone app – that the food is still good. This system could also give a specific guess about how many days are left before it goes bad.

Grell says that if stores don’t have to throw away things just because they’re past their use-by date, even if they’re still good to eat, it could really help stores manage their stock better and stop wasting so much.