Six Hours a Week Botsitting: German Companies Struggle to Turn AI into Productivity
Published on 07/03/2026 at 18:46 | Redaktion boerse-global.de
A wave of artificial-intelligence adoption is sweeping through German mid-sized businesses, but the promised productivity gains remain stubbornly out of reach. According to a survey by the National Bureau of Economic Research (NBER) covering nearly 6,000 executives in Germany, the United States, Britain and Australia, 69 percent of companies now use AI actively. Yet 89 percent of those surveyed report zero positive impact on productivity.
The main culprit, researchers say, is a lack of training. More than half of all employees have never received any instruction on how to use AI tools effectively. This mismatch between technology and skills has created a costly new routine: "Botsitting." A separate Workday study of 200 German firms with over 500 workers found that employees spend an average of 6.4 hours per week monitoring and correcting unreliable AI outputs. While 74 percent of staff consider AI generally useful, only 51 percent say it has actually boosted their own productivity.
New regulations add to the burden
Beyond technical teething problems, German small and medium-sized enterprises (SMEs) face a thicket of new compliance obligations. Since January 2025, mandatory e-invoicing applies in B2B transactions. Since February 2025, parts of the European Union's AI Act – including bans on certain AI practices and a requirement for AI literacy among personnel – have come into force. Workplace time tracking remains another regulatory flashpoint. Germany's Federal Labour Court ruled in September 2022 that employers must document the start, end, duration and overtime of each working day. A planned reform would make electronic time tracking compulsory, with exemptions for very small businesses of fewer than ten employees still being debated. Violations can carry fines of up to €30,000.
When compliance demands are piling up, a missing document can lead to expensive penalties. A free toolkit offers ready-to-use risk assessments and checklists to help UK employers stay on the right side of the Health & Safety at Work Act. Get your free Health & Safety at Work Act 1974 Toolkit
Deep integration remains elusive
Many companies are still running AI in isolation rather than weaving it into their core operations. A CANCOM/techconsult survey from early 2026 shows that 76 percent of mid-sized firms use AI productively, but only 26 percent have fully integrated it into critical systems. Nearly half of the companies operate AI in separate silos. IT service providers are now pushing for deeper connections. SAP, for example, plans to introduce more than 200 AI agents into its business platforms by the third quarter of 2026. Dropbox and M-Files released mid-year updates featuring context-aware AI integrations that automate document management. In manufacturing, government-funded programmes are developing edge-AI solutions that analyse data directly at the machine. Projects such as Easy and Edna focus on predictive wear detection and CO? tracking on the factory floor.
Process before technology
Consultants advising SMEs stress that companies should first map their existing supply chains before deploying AI. One client project was able to reduce weekly inspection time from 30 hours to under five hours after streamlining workflows. Meanwhile, the Zurich University of Applied Sciences (ZHAW) is helping SMEs build Responsible AI strategies. Clear governance structures, researchers argue, can lower compliance costs and build trust in the technology. The lesson from all these findings is blunt: AI alone will not deliver the productivity lift that German business hopes for – what it needs are skilled employees and thoughtfully redesigned processes.
Disclaimer regarding our articles: No investment advice, no buy or sell recommendation. Information on prices, companies, and markets is provided without guarantee; changes are possible at any time. Stock market transactions can lead to substantial losses. Our articles are created and reviewed in whole or in part automatically with the support of AI.
