en de

Online Magazine

AI prevents food waste

Every year, goods worth around 65,000 euros are thrown away in German bakeries. That doesn't have to be the case, says the start-up foodforecast: With the help of AI-supported sales forecasts, it wants to prevent food waste before it even occurs.

by Tobias Imbach

How many croissants will we sell next weekend? Will we sell more in our old town branch tomorrow than in the one in the industrial area? And how many fewer baguettes do we have to order on a rainy day? Bakery employees have to deal with these and similar difficult questions every day.

On the one hand, being able to estimate how much will be sold on a given day is important for the economic success of a bakery. On the other hand, food waste also has a significant impact on the climate.

That is why the start-up foodforecast wants to support bakeries in avoiding food waste: With the help of AI, they create an individual sales forecast for each individual shop.

In the podcast, foodforecast CEO Justus Lauten explains what role weather data plays in this. And where their solution could also be applied.


  • Foodwaste accounts for about 7% of greenhouse gas emissions worldwide. This connection with climate protection motivated me personally to do something about it.
  • From an economic point of view, foodwaste is also an absurdity: every year, a single bakery shop throws away goods worth around 65,000 euros.
  • The advantage of AI is that it prevents food waste before it occurs: Based on a forecast, you produce or order exactly as much as is then sold.
  • From the merchandise management and cash register systems of a bakery, we know exactly what was sold at what price in the last few years. Using this data as well as weather data and information on public holidays, our model learns to create a forecast for each individual shop.
  • We included weather data based on customer feedback: They say that as soon as the temperature rises above 30 degrees, or when it rains heavily, for example, sales generally plummet.
  • Another pattern that emerged was that during Corona, the shops in the inner cities were much emptier than usual, and those in the suburbs were more frequented – presumably because people were working from home.
  • Foodwaste is one thing – but equally, our AI model is designed to prevent a bakery from selling out too early in key product categories.
  • Our AI model does more than simply averaging the latest sales figures: It learns to make forecasts on its own and dynamically adjusts them to changes.
  • Sometimes the AI's recommended order quantity for a product doesn't match what employees would estimate. But most of the time, the AI is ahead because it is data-driven.
  • We are working along the supply chain of our customers: next we want to find out whether it would help mills to know how many kilograms of flour a baker will need in the next two weeks.

What else can AI do?

For example, monitor armed conflicts around the globe.
Read the interview about it here.

Help visually impaired people find their way around.
Fore more information, listen to this podcast.

Detect breast cancer earlier.
Find out more in this article.

Listen to this podcast on another platform:




Justus Lauten

foodforecast, Köln

Justus Lauten's original idea was to help SMEs become more sustainable. When he sent his idea of an AI-supported sales forecast to several companies, bakeries in particular came forward who also knew the problem of high product rejects. This laid the foundation for foodforecast.

Before founding the start-up, Lauten worked as CTO for an e-commerce website. He has a Master of Science in Computer Science from the University of Aachen.

The next round of financing is under way: If you are interested in the project, find out more here.


Data analytics Machine learning

7 habits to shorten the time-to-value in process mining
AI in medicine Data analytics

The smart drinking cup
AI in business Data analytics Machine learning

How can banks become truly AI-driven?