Why the Mittelstand Should Not Operate AI Itself

15 June 2026 · 5 min read · AI in the Mittelstand

88 per cent of companies use AI. Only 39 per cent see a measurable effect on their bottom line. That is the finding of the latest State of AI Report by McKinsey. The gap between adoption and impact is enormous. And in the German Mittelstand, it is growing even wider.

The Mittelstand is falling behind

The Horváth study from January 2026 provides the numbers. Mid-market companies reduced their AI investments in 2025 to 0.35 per cent of revenue. The previous year it was still 0.41 per cent. The overall market rose to 0.5 per cent over the same period. The Mittelstand is investing 30 per cent less than the average.

The Deloitte State of AI Report 2026 confirms this. 42 per cent of the companies surveyed completely abandoned at least one AI project in 2025. The average sunk cost per cancelled project was 7.2 million dollars.

Why AI projects fail

Not because of budget. According to RAND Corporation, 80.3 per cent of all AI projects miss their original business case. The reasons: lack of internal know-how. No usable data foundation. No clear process for embedding AI into day-to-day operations.

Gartner forecasts that 60 per cent of all AI projects without AI-ready data will be discontinued by 2026. Only 28 per cent of AI applications fully achieve their ROI targets.

ManpowerGroup Talent Shortage Survey 2026: AI competence is for the first time the hardest-to-find qualification worldwide. 72 per cent of employers report difficulties. In Germany, around 70 per cent of AI positions could remain unfilled by 2027.

The alternative is not "no AI"

The alternative is "use AI without having to operate AI." Managed AI instead of building your own.

At lupenRhein, we have developed four AI Adapters for exactly this purpose:

AI Adapter Sales: Lead generation with AI. Sales teams with AI support increase their pipeline by a factor of 3 and reduce acquisition costs by up to 65 per cent. 83 per cent report revenue growth. The AI Adapter handles the entire chain from contact identification to a meeting in the calendar.

AI Adapter Marketing: According to HubSpot, AI reduces lead qualification time by 30 per cent and increases lead accuracy by up to 77 per cent. Campaign logic, content production in the brand voice, social publishing, performance tracking.

AI Adapter SEO/GEO: Classic SEO plus Generative Engine Optimization. Visibility on Google and in AI answer engines such as ChatGPT, Perplexity and Google AI Overviews.

AI Adapter Communications: PR, thought leadership, internal communications, stakeholder dialogue. Final approval always remains with the human.

Use rather than operate – the principle

Development, operation and optimisation are handled by lupenRhein. The client's team does not need to learn anything new. A web front end for briefings, approvals and reporting. The Deloitte study confirms: 74 per cent expect revenue growth through AI, but only 20 per cent are already achieving it.

A word on honesty

AI will not triple your sales overnight. It will not replace your marketing department. What AI can do: systematically perform what your team cannot scale. Every single day.