Knowledge: AI in mid-sized industry, clear and sourced
Articles on the questions that come up before any AI rollout: what it costs, what the law says, how to measure the value. Every article is dated and names its sources.
- Runs on your server
- Reads only, never writes
- GDPR ready
- First result in weeks
In short
The articles give orientation, they are not legal advice. Market figures come from KfW and DIHK publications, prices from vendor pages, legal points from the statutes.
What does AI cost in mid-sized industry? All cost items, calculated honestly
The licence is rarely the biggest cost item. An AI solution in mid-sized industry costs implementation, operation, model usage, infrastructure and, above all, the time …
Learn more →AI and GDPR
When a language model is used, the content of a request goes to the model provider, even if the database stays in-house. For personal data this therefore requires a …
Learn more →EU AI Act in mid-sized industry
The AI Act (Regulation (EU) 2024/1689) applies in stages. Since February 2025, prohibitions and the AI literacy obligation have applied, and since August 2025 …
Learn more →Introducing AI with the works council, not against it
The works council must be informed when AI is being planned (§ 90(1) no. 3 of the Works Constitution Act, BetrVG), may call in an expert to assess it (§ 80(3)) and has a …
Learn more →Data masking for AI
Masking replaces or removes names and other attributes before data is used for development and analysis. It reduces risk significantly but does not automatically make …
Learn more →On-premise AI or cloud
“On-premise AI” can mean two things: the platform runs in your own premises, or the language model runs there too. With Onagero, the platform with your data, tools and …
Learn more →Setting up an AI pilot project properly
An AI pilot succeeds when it is settled beforehand: which workflow, who owns it, who releases the budget, which data is reachable, how success is measured, what is …
Learn more →Measuring AI value
Measure the same task before and after, under comparable conditions. Value freed-up time at a realistic share, because free hours are not automatically saved salaries. …
Learn more →AI in mid-sized industry 2026
According to KfW (February 2026), around 20% of mid-sized companies used AI in the years 2022 to 2024. According to the KfW Digitalisation Report, the share of companies …
Learn more →
Start with one problem.
Keep the capability.
No AI transformation. One real process, solved on your data in a few weeks. Your team solves the next one itself.
Common questions
Who writes the articles?
The Onagero editorial team, meaning the people who build the tools. Every article is dated and names its sources.
Are the articles legal advice?
No. They put things in context. For your specific situation, involve your data protection team, your works council and, where appropriate, legal counsel.
How current is the information?
Every article names its date. We update articles when law or prices change, and then change the date too.
Glossary
Getting started
Which problem would
your team solve first?
Show us a real process. We will show you how it could look with Onagero.
Book a demo →