The sidelining of “Made in Germany”
The German economy, long admired worldwide for “Made in Germany,” is sidelined. The automotive industry, its largest and most visible sector, has absorbed year after year of revenue losses. Even Volkswagen, the world’s largest automaker, faces acute distress.
The result is paralysis and shock across the industry. Strategies are discarded. Long-term trends such as electromobility and autonomous driving are subordinated to familiar high-displacement combustion engines.
The same pattern appears in other sectors, but automotive remains the clearest example of decline: a lack of strategy, products that no longer sell, and costly, overcomplex personnel and decision structures. Faster, more innovative international competitors operating under fewer regulatory burdens, and artificial intelligence itself, are eroding these advantages.
In boardrooms, the question is how a company can still innovate under these conditions, and how AI can move faster and reduce costs.
The AI-enhanced trap
This pressure narrows the field of view to an “AI-Enhanced” perspective: how to accelerate one process or another with AI tools. A recent McKinsey survey illustrates the outcome. Eighty percent of respondents report that AI has improved their individual productivity, and half say it helps them make better decisions. At the enterprise level, however, only 37 percent attribute any EBIT impact to AI – the same share as the prior year.
High performers who credit AI with at least 5 percent of EBIT remain a flat 6 percent. Roughly one in five organizations report that AI-related operating costs already constrain further use. Employees gain; the firm does not, and in some cases costs rise.
The AI productivity paradox is the contemporary version of Robert Solow’s 1987 observation: you can see AI everywhere – in tools, individual task speeds, and executive announcements – except in the productivity numbers and the P&L.
The shortfall is structural rather than technical. Time saved is absorbed by verification work (“botsitting”), coordination friction, and the simple failure to redesign workflows, reallocate capacity, or change the operating model. Bolting tools onto a human-centric process architecture therefore produces local abundance without organizational abundance – the exact trap that an AI-native redesign, built as an independent edge twin with humans placed deliberately above the loop, is meant to escape.
The apparent dilemma
For the German automotive industry, the consequences can be stated directly:
- It innovates little and concentrates on cost reduction.
- To become viable, it must innovate rapidly.
- If it continues on its current path, it will not survive.
The dilemma is only apparent. It dissolves when it shifts from an AI-enhanced to an AI-native perspective. AI is no longer treated as a convenient tool applied at scattered points. It is used systematically to innovate faster, more effectively, more efficiently, and more cheaply.
Why the core cannot become AI-native
AI-native thinking and practice do not arise inside the core enterprise. Adding AI after the fact to existing processes therefore produces little structural change.
The core needs an entrepreneurial decision to establish a new unit, separate from and independent of the core company, designed from the outset on AI-native principles. This unit must act autonomously and not depend on the mothership’s processes.
AI-native versus AI-enhanced
Innovation isn’t about technology, or even ideas – it is about finding novel solutions to important problems. Current enterprise AI efforts fail when they simply bolt new tools onto old workflow architecture. A human-centric organization passes work from person to person. An AI-native organization routes work through intelligence layers and places humans deliberately above the loop: setting constraints, validating outcomes, and handling exceptions.
The Edge Twin and the destination
The practical route is to build an AI-native edge twin at the organizational boundary, prove it on real workflows, and migrate work onto it as it outperforms the core. The edge twin must know its destination. Backcasting – defining the future state and working backward – precedes any transformation roadmap.
The evolut strategy is the transformation method. This is a category change. KPMG’s Adaptability Index reaches a parallel conclusion from the outside: most large organizational structures were built for information scarcity and now misfire under abundance. Nearly two-thirds of executives report using data and analytics more, yet fewer than half say decisions are faster or clearer. C-suites lack the vocabulary and frameworks to redesign them.
The safety architecture
Every cycle of recursive improvement must run inside a “govern-assure” control plane. Version and test prompt changes against compliance baselines before deployment. Models that degrade on the evaluation suite are rolled back automatically.
No agent-generated optimization is released without meeting the criteria defined in its structural specification. The compounding advantage is real, but it materializes only under strong, programmatic governance.
What the AI-native path delivers
Choosing the AI-native path dissolves the dilemma. Development cycles shorten dramatically. Innovative products can be produced cheaply. Processes are slimmed. Multiple innovation streams restart. Companies regain a trajectory that is viable beyond the next round of cost-cutting.