Skip to main content
    KI-Beratung Deutschland
    🇩🇪DE🇬🇧EN
    AI in SMEs

    AI in German SMEs in 2026: progress and structural barriers

    Christoph Herting9 min read

    German SMEs are at a turning point in 2026. Large corporations have long established dedicated AI departments, while many companies with 50 to 500 employees are still deciding where and how artificial intelligence can be useful. The available data presents a mixed picture, with clear progress in some industries and significant catching up to do in others.

    Current figures: where do SMEs actually stand?

    We assess the achievable impact in a pilot using jointly defined metrics. A specific financial outcome is not guaranteed.

    Most companies remain far from a systematic AI strategy spanning several departments or their business model. Only twelve percent of surveyed businesses reported a formal AI roadmap extending beyond individual projects. In practice, a bottom-up approach still dominates: individual departments test tools without coordination across the organisation.

    The differences between industries are striking. Mechanical engineering and automotive suppliers are furthest ahead, with AI for predictive maintenance and quality control already part of their standard toolkit. Retail and logistics are expanding demand forecasting and route optimisation. Skilled trades, construction and parts of the services sector remain considerably more cautious.

    The three main barriers: skills, data and investment certainty

    Why do many SMEs hesitate when capable AI tools are more available than ever? The reasons are organisational more than technical.

    First, the skills shortage. AI expertise remains scarce and expensive. Large companies and technology groups offer salaries that an owner-managed company in the Sauerland region simply cannot match. Many SMEs therefore use external consultants, which makes sense when those consultants build lasting capability rather than dependence.

    Second, data quality. AI models are only as good as the data they use. In many companies, relevant information is scattered across spreadsheets, ERP systems and email inboxes. Consolidating this data often takes more work than the AI implementation itself, but without it every AI project falls short of its potential.

    Third, investment uncertainty. Many managing directors wonder whether the technology will still be relevant a year later or whether they are choosing the wrong approach. This is understandable, but an incremental approach can reduce uncertainty: a clearly defined pilot whose ROI can be demonstrated within three to six months limits risk and creates a basis for further investment decisions.

    Comparing industries: who benefits most?

    There is no single answer to which industries gain the most from AI. The type of process being automated matters more than the industry. As a general principle, a higher proportion of repetitive, rule-based tasks brings a faster payback.

    We assess the achievable impact in a pilot using jointly defined metrics. A specific financial outcome is not guaranteed.

    Acceptance of AI-assisted documentation is growing in healthcare. Medical practices and medical care centres report saving up to two hours a day when preparing findings and clinical letters, leaving those hours available for patient care.

    What this means for competitiveness

    A company's lag in adoption can become a competitive problem in the medium term. AI technologies are becoming more efficient and less expensive. Companies investing now build a cumulative advantage each quarter: better models, richer data and leaner processes.

    Companies waiting for competitors to move first face two challenges: their own transformation effort and the lead already established by competitors. In price-sensitive industries, that difference can affect market share.

    Rushing into projects is also counterproductive. Not every AI application is useful and not every process suits automation. The key question is where to use AI. A careful potential analysis objectively assessing costs, benefits and risks is the first, often underestimated step.

    A pragmatic starting point

    German SMEs will not become AI leaders overnight in 2026, and they do not need to. What matters is a pragmatic start: a concrete use case, a measurable result and gradual expansion. Companies following this path almost universally report positive experiences, because the first success creates organisational readiness for further steps, even when the technology is imperfect.

    The greatest danger is making no decision and remaining stuck in analysis. Markets do not wait.