Best Practice Digitalization : AI in the warehouse: expectation vs. reality

At the time of the survey, 15 percent of companies were already using AI productively in their warehouses. Another 41 percent announced their intention to invest in corresponding solutions by 2028. The focus is on warehouse optimization and inventory management as the most important areas of application. At the same time, several obstacles such as cost risks, employee acceptance, and technical challenges were identified.

You can read more about the sample and methodology of the PSI study in the press release PSI Industry Study Shows Trend Toward Artificial Intelligence in Warehouse Logistics.

In the summer of 2026, it is clear that the strong interest in AI has remained. What is still missing is the pace of implementation. A noticeable gap remains between the intention to invest and productive use. The use of AI in logistics mostly takes place outside of the actual warehouse processes – for example, for evaluating information or analyzing documents. Many companies have significant reservations about integrating AI into business-critical processes. And these concerns are understandable. AI cannot be a black box with a vague purpose. It must operate within clear boundaries and deliver comprehensible results.

Solutions that make it into productive logistics operations usually provide support in individual, clearly defined application areas: in picking, inventory analysis, or recurring decisions. What becomes productive is what meets three conditions: it fits seamlessly into existing processes, it delivers a quickly measurable benefit, and it remains understandable to the user.

The maturity of an AI solution only becomes apparent when an innovation project turns into a ready-to-use product: calculable in cost, immediately available, and with a directly tangible benefit for the user. PSI is developing its AI portfolio in precisely this direction. A first result of this strategy is Batch AI: The AI module is now part of the PSIwms standard. In practical use, it has helped customers reduce picking routes by around 30 percent and increase picking efficiency by over 20 percent. This is just the beginning: Following the model of Batch AI, PSI will develop further AI modules that offer concrete advantages in very practical application areas.

You can learn more about PSI's AI strategy in the blog article New AI modules for improved warehouse processes.

The interim assessment after one year is relatively calm – and that is good news. AI will only become standard in the warehouse through comprehensible, clearly defined solutions with measurable benefits. Those who see AI as a toolbox rather than a complete package will achieve productive use more quickly. PSI is continuously developing its portfolio in exactly this direction.