AI in the German Mittelstand: Strategy Gap, Facts, and a Framework for Action

Miguel Bebensee

February 13, 2026

43 % of German mid-sized companies have no concrete AI plans—while 91 % of large enterprises already consider AI business-critical. This article analyses the current data landscape using the DMB/Salesforce Mittelstand AI Index, the 2025 KPMG study, and other sources, presents ROI benchmarks (18–35 % productivity gains), and provides a five-step framework for AI strategy development. Supplemented by regulatory requirements under the EU AI Act and NIS2.

Key Takeaways at a Glance

Germany’s Mittelstand is reaching an AI inflection point: while the use of artificial intelligence has nearly doubled within a single year, 43 % of mid-sized companies still lack any concrete AI plans. This strategy gap is becoming an existential risk, as the divide between AI leaders and laggards is widening rapidly. The 2025 KPMG study shows that 91 % of large enterprises already consider AI business-critical and are ramping up budgets significantly—the Mittelstand risks being left behind. At the same time, recent studies demonstrate that companies with a strategic approach to AI achieve productivity gains of 18–35 %, with AI investments often paying for themselves within 6–18 months. The good news: proven frameworks and advisory services make getting started easier than ever before.

The 2025 Mittelstand AI Index Paints a Divided Picture

The key data source for the strategy gap is the “KI-Index Mittelstand” (Mittelstand AI Index), published by the German Mittelstand Association (DMB) and Salesforce in February 2025. The study surveyed 526 mid-sized companies with up to 500 employees (fieldwork period: November–December 2024). The core findings:

33.1 % of Mittelstand companies are already using AI—of these, 9.5 % have fully implemented AI and around 24 % are in the pilot phase. A further 24.9 % plan to adopt AI within the next twelve months. Yet 43 % have no concrete AI plans for 2025—this is the widely cited strategy gap. The study’s original wording precisely refers to “no concrete AI plans for this year,” which is frequently shortened in media coverage to “no AI strategy.”

Among AI users, generative AI dominates at 72.6 %, while predictive AI accounts for only 12 % and AI agents for roughly 10 %. The biggest barrier is a lack of background knowledge about specific use cases (27.4 %), followed by regulatory uncertainty (20.6 %) and insufficient technical expertise (13.7 %).

Marc S. Tenbieg, Managing Director of the DMB, summarised: “The results clearly show that artificial intelligence represents both an enormous challenge and a tremendous opportunity for the German Mittelstand. Knowledge and skills gaps, along with investment uncertainty, are holding back the innovative capacity of many companies.” Alexander Wallner, CEO of Salesforce Germany, offered a more positive assessment: “One third active AI users is a good starting point. AI adoption is a once-in-a-century opportunity for the German Mittelstand.”

KPMG Study 2025: From Nice-to-Have to Must-Have

The KPMG study “Generative AI in the German Economy 2025” (published June 2025, 653 decision-makers across 18 industries) provides the large enterprise perspective—and thus the benchmark against which the Mittelstand must measure itself. The momentum is striking: the share of companies rating AI as important for their business model jumped from 55 % (2024) to 91 % (2025). The share with an AI strategy doubled from 31 % to 69 %, with a further 28 % actively developing one.

Investment appetite is surging: 82 % plan budget increases (previous year: 53 %), half of them by at least 40 %. Benedikt Höck, Head of AI at KPMG, noted: “Generative AI has evolved from an experimental playground into a strategic imperative.”

The study’s central warning: the gap between AI adopters and non-adopters keeps widening. While 69 % of (predominantly large) companies have an AI strategy, the complementary SME study by maximal.digital (455 SMEs) shows that 68 % of SMEs have no AI strategy at all. In other words, the Mittelstand’s strategy gap is almost a mirror image of the large-enterprise landscape.

It is important to note, however, that the KPMG sample is skewed towards large enterprises (77 % of respondents from companies with ≥500 employees). The high adoption figures primarily reflect the large-enterprise perspective—the reality for the Mittelstand is likely far more sobering.

AI Adoption in Germany: From 20 to 40 Percent in One Year

AI adoption in Germany accelerated dramatically between 2024 and 2025. Across all sources, a consistent picture of growth emerges, even though the absolute figures vary depending on methodology:

The Bitkom study 2025 (604 companies with 20+ employees) reports 36 % AI adoption; the ifo Institute even reaches 40.9 %. For the Mittelstand in the narrower sense, figures are somewhat lower: the HKA/KARL study (517 mid-sized firms with 20–500 employees) finds 40 %, while the Federal Network Agency reports just under 30 % among SMEs. Broken down by company size (Destatis 2024): large companies at 48 %, medium-sized at 28 %, small at 17 %.

Particularly striking is the planning momentum: only 17 % of companies still say AI is “not a topic,” according to Bitkom—down from 52 % in 2023. The Sage study (January 2025) reports that 63 % of AI-using SMEs are achieving direct business improvements.

Leading sectors include advertising/market research (84.3 % per ifo), IT services (73.7 %), automotive (70.4 %), and consulting (53 %). Laggards include construction (25 %), wholesale/logistics (24.1 %), and food manufacturing (21 %).

The most common barriers are remarkably consistent across sources:

Lack of expertise and talent shortage (53–72 % depending on the study)
Regulatory uncertainty, particularly around the EU AI Act (51–53 %)
Insufficient data quality and data silos (53–76 %)
Data protection and security concerns (39–48 %)
Missing AI strategy (only 32 % of SMEs have one)

AI ROI: Between Productivity Boost and Reality Check

The evidence on measurable business value from AI is mixed. On the one hand, impressive figures: the IBM study “The Race for ROI” (September 2025, 3,500 executives) shows that 62 % of German companies report significant efficiency gains from AI. Nearly half expect a return on investment within 12 months. The OECD study on GenAI in SMEs (2024) finds that 65.1 % of GenAI-using SMEs report improved employee performance and 45.2 % achieve cost savings.

On the other hand, the PwC CEO Survey (4,400 CEOs worldwide) reveals that two thirds of German CEOs have yet to see a positive impact from their AI investments. Only 11 % achieved higher revenues; just 2 % managed to both increase revenue and reduce costs. This discrepancy is explained by differing samples and time horizons—many AI projects need time to deliver results.

Macroeconomic forecasts remain optimistic: PwC estimates AI could boost German GDP by 11 % (approx. €430 billion) by 2030. McKinsey puts productivity growth from rapid AI adoption at up to 3 % per year in Europe. Accenture forecasts a 30 % increase in labour productivity by 2035.

Figures specifically relevant to the Mittelstand: companies with a clear AI strategy achieve productivity gains of 18–35 % according to the maximal.digital study. AI-powered quality control reduces defect rates by up to 40 %. A typical chatbot project (€50,000 investment) pays for itself through saved service costs in approximately 8 months. In industries with intensive AI use, productivity growth has quadrupled from 7 % (2018–2022) to 27 % (2018–2024) according to PwC.

Five Steps to an AI Strategy: What Frameworks and Experts Recommend

Several proven frameworks exist for developing an AI strategy in the Mittelstand. The best-documented comes from Plattform Lernende Systeme (supported by BMBF) in collaboration with appliedAI and follows five milestones:

Step 1 – Assess the status quo: Determine digital maturity, evaluate existing data assets, identify automatable processes, and take stock of team competencies. Tools such as the AI Readiness Check from Mittelstand-Digital or the AI Maturity Assessment from appliedAI offer free entry points.

Step 2 – Evaluate AI value: Clarify strategic objectives, define KPIs, prioritise specific use cases, and conduct an initial ROI estimate. An impact-effort matrix helps with prioritisation.

Step 3 – Design the AI strategy: Conduct competitive analysis, make build-or-buy decisions, plan resources, and select technologies. Key insight: 89 % of SMEs use external solutions rather than in-house development.

Step 4 – Launch pilot projects: Start in areas with high existing digitisation, learn iteratively, and adjust strategy based on results. The Detecon roadmap recommends quick wins in the first 0–6 months.

Step 5 – Scale and embed: Roll out successful pilots, drive change management, and optimise continuously. Mittelstand-Digital (Federal Ministry for Economic Affairs) offers 29 centres nationwide and around 100 AI trainers providing free, vendor-neutral support.

Recommended quick wins for getting started, according to industry associations and consultancies: AI chatbots in customer service, automated document processing and email categorisation, AI-driven demand forecasting, generative AI for marketing content creation, and standard tools like Microsoft Copilot or ChatGPT Enterprise as a low-barrier entry point.

The ten most important success factors, synthesised across sources: management commitment (AI is a leadership responsibility), clear strategic goals, starting with 2–3 focused use cases, ensuring data quality (76 % struggle here), active employee engagement and upskilling, early development of AI governance, pilot projects with measurable ROI, involving business units, leveraging external networks, and deliberate build-or-buy decisions.

EU AI Act and NIS2: Two Regulations That Demand Immediate Action

Managing directors in the Mittelstand face a dual regulatory challenge. The EU AI Act has been in force since 1 August 2024 and is being phased in gradually. Since 2 February 2025, the ban on AI systems posing unacceptable risk (social scoring, manipulative AI) and the AI literacy obligation under Article 4 already apply—every company must ensure that employees working with AI are adequately trained. On 2 August 2026, the requirements for high-risk AI and transparency obligations take effect: chatbots must be labelled as AI, and AI-generated content must be made transparent.

The AI Act follows a risk-based approach with four tiers. For most Mittelstand companies acting as deployers (not developers) of AI systems, the obligations are manageable but real: use systems according to instructions, carefully select input data, monitor operations, and retain logs. However, anyone who substantially modifies an AI system or markets it under their own name becomes a provider with significantly more extensive obligations. Penalties are severe: up to €35 million or 7 % of global annual turnover for prohibited practices, with the lower amount applying for SMEs. The supervisory authority in Germany is the Federal Network Agency (Bundesnetzagentur).

The NIS2 Directive (cybersecurity) was transposed into German law on 6 December 2025 with no transition period. It affects companies with 50 or more employees or €10 million or more in annual turnover in 18 designated sectors—including energy, transport, healthcare, but also mechanical engineering and manufacturing. Approximately 29,000–30,000 companies are directly affected; through supply chain requirements, potentially over 200,000 more. The decisive paradigm shift: managing directors are personally liable for their company’s cybersecurity—with their private assets. Simply delegating to the IT department is no longer sufficient. Registration with the Federal Office for Information Security (BSI) must be completed by early March 2026; a 24-hour incident reporting requirement applies immediately.

Both regulatory frameworks overlap: AI systems deployed in critical infrastructure may be classified as high-risk AI and thus fall under both regulations. The most efficient preparation is an integrated management system based on ISO 27001 that covers AI governance and IT security together.

Conclusion: The Next Twelve Months Will Be Decisive

The data is clear: the Mittelstand stands at a crossroads. AI adoption is accelerating exponentially, the divide between leaders and laggards is growing, and regulatory requirements are creating unmistakable time pressure. Three key insights stand out:

First, the absence of strategy is the real risk—not the technology itself. The 43 % without AI plans from the DMB/Salesforce Index face a market in which 91 % of large enterprises rate AI as business-critical. Companies that fail to develop a strategy now will not lose tomorrow, but will systematically lose competitiveness within two to three years.

Second, getting started is easier and more affordable than ever. Standard AI tools such as Microsoft Copilot, ChatGPT Enterprise, or industry-specific SaaS solutions drastically lower the entry barrier. The 29 Mittelstand-Digital centres offer free guidance, and regional networks provide direct access to research and practice.

Third, regulatory compliance is becoming a competitive advantage. Companies that build AI governance now and prepare for the EU AI Act are simultaneously creating the structural foundations for scalable AI projects—and will be preferred as trustworthy partners in supply chains that increasingly require NIS2 conformity. The companies that have their AI strategy in place by 2026 will be the ones shaping their industry by 2028.

Sources

DMB / Salesforce – Mittelstand AI Index (February 2025): mittelstandsbund.de | salesforce.com/de
KPMG – “Generative AI in the German Economy 2025” (June 2025): kpmg.com/de | kpmg.com/de (press release)
maximal.digital – AI Study 2025: AI in Mittelstand and SMEs: maximal.digital
Bitkom – Artificial Intelligence 2025: bitkom.org | bitkom-research.de
ifo Institute – AI Adoption in German Companies: ihk-muenchen.de | marktundmittelstand.de
HKA / KARL Study – AI Use in the German Mittelstand (2025): h-ka.de
Destatis – Use of AI in Companies (2024): destatis.de
Federal Network Agency – Digitalisation Indicators for the Mittelstand: bundesnetzagentur.de
Sage – German Mittelstand Leads Europe in AI Adoption (June 2025): sage.com/de-de
IBM – “The Race for ROI” (October 2025): de.newsroom.ibm.com
OECD – Generative AI and the SME Workforce (2024): oecd.org
PwC – CEO Survey / AI and Labour Productivity: forum-institut.de | de.statista.com
McKinsey – GenAI & Future of Work (May 2024): mckinsey.de
IW Köln – AI as a Competitive Factor (2025): iwkoeln.de
Plattform Lernende Systeme (BMBF) – AI Strategy Framework: plattform-lernende-systeme.de | Practical Tips
Mittelstand-Digital (BMWK) – AI Trainers and Guides: mittelstand-digital.de | Steps to Integration
appliedAI – AI Maturity Assessment: appliedai.de
bidt – Topic Monitor: AI in the German Mittelstand 2025: bidt.digital
EU AI Act – Legal Text and Timeline: ai-act-law.eu | alexanderthamm.com | activemind.legal
Mittelstand-Digital – Brief Study on the AI Regulation: mittelstand-digital.de (PDF)
NIS2 Implementation Act – Overview and Liability: haufe.de | twobirds.com | orbit.de
DSGV – AI and SMEs (2025): dsgv.de (PDF)