Imagine running a company that makes towels and bathrobes, trying to guess how many to make next month. Most businesses still use Excel spreadsheets and rely on whatever their most experienced employee remembers from years past. Now, researchers in Germany have built an AI tool that makes those guesses much better — and it all started in a small town in eastern Germany.
Fraunhofer IWU, a research institute in Upper Lusatia, Germany, partnered with frotana Textil GmbH & Co. KG, a longtime towel and home textiles maker behind the MÖVE brand. The researchers created a computer program that studies past sales numbers and predicts what customers will buy in the future. The AI looks for patterns, like how towel sales spike every spring or dip around October.
The results were surprising. The tool explained 82.7% of the ups and downs in sales — a strong score for any prediction system. When the company sells an average of 340 towels per month, the forecast is off by only about 38 towels, or roughly 9%. That kind of accuracy helps factory managers plan orders, schedule workers, and avoid wasting materials. Even with just four years of past data and no information about different sales regions or promotional campaigns, the system still performed well.
Before this, frotana's planning felt like guesswork. Some businesses recreate the same Excel spreadsheets every month for thousands of products. Others still use handwritten lists. When skilled workers leave or call in sick, that institutional knowledge walks out the door with them.
The new tool does not replace human workers, though. Managers can still review the AI's predictions and add their own expertise. If someone knows that a local festival tends to boost sales in a certain region, they can adjust the numbers. Researchers call this keeping people at the center — humans and machines working together rather than one replacing the other.
Looking ahead, the team hopes to connect the forecasting tool directly to the factory floor. That way, when the AI predicts a busy month ahead, production machines could automatically adjust batch sizes and schedules. For a region like Upper Lusatia, where small textile companies have faced global competition for decades, this kind of smart planning could help them stay afloat and compete on quality instead of just price.
For frotana, it means fewer surprises, smoother operations, and a business that's ready for whatever the next season brings.