SAP Study Shows: AI Investments Rise, but Governance Remains a Weak Point

What’s It About?

Companies in Germany are pouring considerable sums into artificial intelligence – on average just under 40 million dollars. This is confirmed by a new SAP survey on the value of AI technologies. At the same time, the study reveals a critical gap: governance structures are not keeping pace with the speed of technological adoption. While nine out of ten respondents assume that so-called agentic AI will fundamentally transform their organizations, control mechanisms and clear processes for the use of autonomous AI systems are lacking in many places.

Background & Context

Expectations of the profitability of AI investments have risen: the expected return on investment climbed from 17 percent in the previous year to 24 percent currently. This signals growing confidence in the technology. However, only a fraction of companies – specifically four percent – see themselves as fully equipped for the use of AI agents that can independently take on tasks.

Particularly problematic: only eleven percent of the executives surveyed rate their existing control mechanisms as sufficient. More than half of organizations have no established processes at all for controlling autonomous AI systems. This discrepancy between willingness to invest and operational preparedness could become a risk factor. Companies fear that they are rolling out AI agents faster than they can effectively monitor them. Additional hurdles are poor data quality and the challenge of imparting the necessary competencies to employees.

In practice, application fields are already emerging: AI agents are used, for example, in accounting and production planning, where they achieve measurable time savings. In this context, the principle of “human in the loop” – the involvement of human decision-makers – remains an essential factor for responsible implementation.

What Does This Mean?

The gap between AI investments and governance maturity represents a strategic risk – companies must build control structures in parallel with the technology.
The clear rise in ROI expectations from 17 to 24 percent shows that AI is increasingly perceived as a profitable investment, no longer merely as an experimental technology.
Only four percent full readiness alongside 90 percent transformation expectation signals a dangerous gap between vision and operational reality.
Missing governance processes at more than half of companies could lead to compliance problems, security risks, and inefficient AI use.
The successful pilot projects in accounting and production provide blueprints for other areas – provided that the necessary framework conditions are created.