The AI Cost Trap: Why the Added Value of Billion-Dollar Investments Remains Elusive

Study Download

• Value gap persists for companies despite billion-dollar AI investments.
• Flying blind on success measurement: Most companies rely on intuition rather than data-driven analysis.
• DACH region remains an observer in generative AI and AI agents.

Munich, March 2026: Companies worldwide are investing more money in artificial intelligence every year, but the economic benefits have so far fallen short of expectations. According to the latest Roland Berger study “The AI Value Gap,” nearly 90% of the more than 200 companies surveyed across various regions (Europe, Japan, USA) report that the financial returns from their AI projects lag behind their investment levels. Only a small portion of companies have so far realized concrete economic benefits.

"The value creation gap in AI usage does not exist due to a lack of ambition, capital, or access to technology. Rather, it is because companies treat AI as a purchasable project instead of a capability to be developed." – Maria Mikhaylenko, Global Managing Director for Tech and Knowledge & CTO at Roland Berger.

The Roland Berger study identifies the biggest hurdles in implementing generative AI and AI agents as superficial measurement of success, conflicting objectives, and technical obstacles.

Superficial Measurement of Success:
The majority of companies are still flying blind when it comes to measuring the success of their AI initiatives: 63% of surveyed firms rely solely on one-off measurements or gut feeling instead of continuous, data-driven monitoring. Only a quarter of companies measure the returns of their AI tools automatically and continuously during operation, rather than just once after project launch. As a result, many companies do not know which projects actually generate value and which only generate costs.

Conflicting Objectives:
Many of the companies surveyed dream of maximum autonomy while simultaneously wanting maximum security. This resulting conflict of objectives leads to a decision-making paralysis. At the same time, so-called "shadow AI" solutions are increasingly emerging: departments and individual teams develop AI applications outside of central company IT and governance. This leads to uncontrolled, hard-to-integrate systems and increased risks for security and transparency.

Technical Obstacles:
While many companies purchase modern AI tools, integration remains superficial. Existing IT systems are only superficially extended by new AI solutions, instead of building an interconnected ecosystem. This leads to problems, especially when introducing AI agents. They can think logically but cannot reliably operate within company systems. There is also a lack of technical skills: two-thirds of companies that are stagnating in AI integration cited a lack of technical know-how as a significant hurdle.

"Companies buy Ferrari technology but run it with go-kart engines. AI that is only superficially layered over outdated systems destroys its value." – Maria Mikhaylenko, Global Managing Director for Tech and Knowledge & CTO of Roland Berger.

DACH Region Remains Observer – Japan as a Role Model:
The study also looks at Japan and shows how things can be done better. There, the proportion of companies already deriving economic benefit from AI is more than four times higher than in the DACH region. Japanese companies rely on clear governance, company-wide integration, and continuous value measurement. According to the study, the DACH region remains in an observer role: it prioritizes security and accuracy over autonomy and is still piloting individual AI projects. This reflects its rather cautious approach, but in the long-term global AI race, it could lead to being left behind.

About the Study:
The study is based on a survey of 203 executives with technology-related mandates and direct oversight of or involvement in AI initiatives. The sample covers Europe, the USA, and Japan with the following geographic distribution: DACH region (Germany, Austria, Switzerland) 35%, France 15%, United Kingdom 15%, USA 20%, and Japan 15%. Respondents represent various industries, including: consumer goods, retail and logistics; healthcare and energy; financial services; industry and automotive; as well as technology, media, and telecommunications.