- 57% of companies in Germany use AI in 2026.
- Only 3% of users make fairly strong use of its potential.
- Start with a clearly defined pilot and measure the impact.
AI has arrived in German mid-sized companies, but its potential is far from exhausted. According to Bitkom, a majority of companies in Germany, 57 percent, use AI in 2026 for the first time. Yet none of them say they use its full potential, and only 3 percent think they are making fairly strong use of it. Companies that start in a structured way now can clearly pull ahead.
What the figures show
Bitkom's September 2026 survey paints a clear picture: 91 percent of companies consider AI the most important technology of the future, and 80 percent see it mainly as an opportunity for their own business. Two thirds of users have improved their competitive position with AI, and 45 percent of all companies are spending more on AI in 2026 than the year before.
The flip side: 66 percent of companies rate their employees' AI skills as low. Technology is not the problem; the question is where and how to use it well.
Where AI delivers most in mid-sized companies
- Administration and accounting: capture invoices, delivery notes and orders automatically.
- Customer service: answer recurring questions with a chatbot, around the clock.
- Sales: prepare quotes from requests and qualify leads.
- Knowledge: make manuals, policies and project documents searchable by question.
- Production: analyse quality data and detect faults earlier.
The three most common mistakes when starting
- Starting too big: a company-wide AI programme without an early win quickly loses support.
- Choosing the wrong process: not every workflow is suitable. Best are frequent, rule-based tasks with unstructured data like emails or PDFs.
- Forgetting the people: without training and clear rules, teams don't use new tools, or use them wrongly.
How to start right
- Map processes: where does your team spend time typing, searching or giving repetitive answers?
- Prioritise: rate each use case by value, effort and risk.
- Run a pilot: one clearly defined case, with measurement before and after.
- Train your team: AI literacy has also been mandatory under Art. 4 AI Act since February 2025.
- Scale: roll out what works to other areas.
Not sure where to begin? In an AI opportunity assessment we find the right first use case together. Or request our free AI Scan.
Frequently asked questions
How many companies in Germany use AI?
According to Bitkom, a majority of companies in Germany, 57 percent, use AI in 2026 for the first time.
Where should a mid-sized company start?
With a clearly defined pilot in a frequent process, such as document capture or customer enquiries, and measurement before and after.
Is AI too expensive for small companies?
Not necessarily. Many applications build on existing systems. What matters is a use case that pays off quickly.