Artificial intelligence is often described as a key technology for the industrial future. But how present is it already in practice? And what are the specific requirements for companies looking to introduce it?

In order to obtain a realistic picture of the mood, roeren interviewed experts from various industries and company sizes. The aim was to gain an assessment of the use, potential, obstacles and prerequisites for the use of AI in the production environment from the perspective of operational managers.

Applications: initial experience, but limited range

The survey shows: The majority of companies have only used AI to a limited extent so far. They mainly mentioned pilot projects and niche solutions – for example in quality control, documentation or support from AI assistants such as Microsoft Copilot. It is particularly striking that many applications are used in the environment rather than directly in production itself.

Where companies see potential

Despite the restrained use, experts see substantial potential – especially in the following areas:

• Process automation, e.g. through autonomous transport systems
• Production planning and control, including for better forecasting and faster response to disruptions
• Quality control through image processing and data-based error detection
• Maintenance, for example via predictive maintenance
• Process optimization through real-time data analysis

The survey clearly shows that the fields of application extend across the entire production chain and that there is a great need for solutions to increase efficiency, flexibility and error prevention.

Why implementation is nevertheless faltering

Despite the potential identified, several factors are hampering implementation. The most frequently mentioned were

• lack of internal competencies,
• lack of human resources,
• unclear approaches and a lack of strategies,
• Concerns about data protection, data quality and system reliability.

Data availability as a central prerequisite

Another key finding: the quality and availability of digital data remains a weak point. The respondents rated their own data situation with an average of 5.1 out of 10 points. Without structured, accessible and high-quality data, many AI projects remain purely hypothetical or cannot be scaled up.

Responsibility and structure: Who drives AI internally?

The organizational anchoring of AI is still inconsistent. In most cases, it is located in IT. Only rarely are there independent AI departments or integration into production-related areas such as LEAN or Operational Excellence. The results indicate that a lack of responsibilities makes it difficult to transfer AI into day-to-day operations.

What companies really need

When asked about the need for support, there was a desire for pragmatic approaches: Further training, standardization, modular solutions and strategic data concepts were mentioned particularly frequently. External support for potential analysis and strategy development is also considered useful.

Conclusion

The results make it clear: There is not a lack of will, but a lack of prerequisites. So far, AI has been used selectively in production. However, widespread use requires significantly more structure, expertise and integration. Companies that make targeted investments in these prerequisites can exploit the potential of AI much better and ensure sustainable value creation along the entire production chain.