German manufacturers turn years of factory data into AI projects

In short: German manufacturing companies are starting to use decades of factory data to build and run AI tools, hoping it helps them compete and cut costs.

What's going on

German industrial companies, including many Mittelstand firms (small and medium-sized businesses, often family-run), are looking at their old manufacturing records as a useful resource for AI. The idea is simple, AI systems learn from examples, like a trainee who improves after reading years of past job reports.

Some companies say they already have a head start because they have collected data for a long time. Herding, a filtration equipment maker with about 400 employees, said it has around 50 years of “experience data,” including technical details from thousands of projects. It is also using AI to read handwritten reports and safety documents so they can be stored and searched digitally.

Others are changing how they work to keep up with faster decision cycles. Safelog, a logistics robot maker, said its day-to-day operations changed more in the past six months than in the previous 10 years. It reorganised around an internal AI platform that gives staff access to “everything the company knows,” and uses it to diagnose production issues and support product development.

The shift is not easy. Many firms did not collect the right data, or they stored it in formats that are hard to use. Cleaning and combining data can take time and money, and Germany still faces wider challenges like high energy and labour costs, heavy paperwork, and relatively low digital adoption.

What to watch

Adoption is still early. A KfW survey found only 1% of Mittelstand companies use AI to automate machinery, and 3% use it for data analysis, even as interest grows. Watch whether more firms invest in turning old records into usable digital data, since that could decide who improves product quality and speeds up development, and who falls behind.

Source: Financial Times