AI Pays Off—But the Data Set Remains the Bottleneck

About seven out of ten companies in Switzerland are now achieving a measurable return on investment from artificial intelligence. This is shown by the latest AI Momentum Index from Dun & Bradstreet. The largest share—48 percent—reports sporadic ROI in individual projects, while another 21 percent is already achieving broad or strong ROI across multiple projects. Nevertheless, only 7 percent of companies consider their data to be fully ready for large-scale AI deployment.

AI Adoption Is Outpacing Data Maturity

The results make it clear: While the adoption of AI is advancing rapidly, the quality of the underlying data is not keeping pace. Gary Kotovets, Chief Data and Analytics Officer at Dun & Bradstreet, sums it up: «The challenge today is that AI adoption has outpaced data maturity. Most organizations are now seeing initial returns from individual pilot projects, but so far this has led to broad, company-wide success for only a few.»

Just under half of Swiss companies are already scaling up AI, firmly embedding it in core processes, or implementing agent-based workflows. In the case of the latter, Switzerland—at just under 8 percent—is actually two percentage points above the global average. There is also no shortage of willingness to invest: Just under 30 percent of Swiss companies expect a sharp increase in their AI spending over the next twelve months—nearly twice as many as the global average.

Data quality remains a critical bottleneck

A closer look at the details reveals where the problem lies: 55 percent of companies describe their data as partially ready, and 22 percent as largely ready for AI-based value creation on a large scale. Just over one in six companies is barely prepared—or not prepared at all—for this. There thus remains a significant gap between the data that enables early AI advancements and the data that can drive scalable, repeatable ROI across the entire company.

Uncertainty also remains high when it comes to risk management: Not even half of companies feel well or very well prepared to reliably identify and manage risks such as bias, data protection, or compliance in connection with AI.

Verified Data as the Key to Agent-Based AI

Kotovets explains what this means in practice: «For AI to make good decisions—whether as part of a team, within a system, or as an independently operating AI agent—it needs a reliable overview of the company’s specific business context. That’s exactly what our identity infrastructure provides: It anchors AI in real, verified company data and gives companies the confidence to let their AI agents work truly autonomously.»

Dun & Bradstreet’s AI Momentum Index shows steady progress in AI adoption and maturity for 2026, even though the data foundation remains a challenge. The third-quarter results suggest that the next phase will require companies to make targeted improvements to their data maturity in order to transform early, unevenly distributed ROI into more consistent returns on an enterprise-wide scale.

Source: www.dnb.com