AI’s 6 Percent Problem [Sponsored]

Nearly 90% of companies now utilize AI in at least one business function, with 44% having scaled it across their entire organization. However, a McKinsey study reveals that only 37% report a measurable impact on EBIT. Only about 6% of companies are considered "high performers," attributing at least 5% of their EBIT to AI.

The challenge lies in accountability. Companies often struggle with data silos, where customer data is scattered across various systems, and revenue figures are maintained in separate spreadsheets. When an AI model queries product availability, it may receive conflicting answers from different sources, leading to errors that need to be addressed later.

Furthermore, ownership issues arise as IT manages the technology, business teams handle use cases, and management expects results without a single person taking full responsibility.

Successful companies address these issues by first improving their data foundation before implementing additional tools. They integrate AI directly into existing systems like CRM or service desks, rather than adding more tools on top. Moreover, they assign one individual to own the outcomes, rather than spreading responsibility among multiple departments.

Notably, the 6% of companies achieving meaningful results with AI share a common trait: they focus on a limited number of high-potential use cases and measure success based on financial outcomes, not just operational metrics. This approach contrasts with the previous year, where only 55% of companies had redesigned their workflows to incorporate AI.

The trend extends beyond business operations to customer interactions. People now engage with AI through various channels, such as AI assistants, social media, and marketplaces. To remain visible, brands must ensure their product data is consistent and machine-readable. This shift in consumer behavior highlights the need for AI to be present throughout the entire customer journey.

These insights were prominent at DMEXCO 2026 in Cologne, where the theme of "Scaling Intelligence" was explored. Over 40,000 attendees and 1,000 speakers gathered to discuss the practicality of AI in the real world. A recurring theme was the importance of having a "human in the lead," emphasizing the need for human oversight and responsibility in AI-driven processes.