Extending the Optimisation Layer<\/p>\n
Inform has long been associated with optimisation in complex, data-driven environments. But as markets become more volatile, optimisation systems are being asked to react faster, incorporate more signals and support more dynamic decision-making.<\/p>\n
That shift is particularly relevant in environments where decisions are interdependent. A change in demand planning may affect inventory, transport capacity, labour allocation or service levels. A recommendation made in one part of the operation can create consequences elsewhere, which makes transparency essential for day-to-day use.<\/p>\n
For Heinrichs, this is where AI in logistics must prove its practical value. \u201cI don\u2019t talk about AI. I talk about explainable AI,\u201d he says. \u201cEverything we do, everything we propose, has an explanation. Otherwise, people don\u2019t trust it.\u201d<\/p>\n
![\"\"](\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/05\/ZEL018-Image-2-1300x729.jpg\")
<\/p>\n
Trust as a Practical Requirement<\/p>\n
In conversations with customers across different industries, he says the same question comes up repeatedly: \u201cWhy did the system pick that option and not another one?\u201d<\/p>\n
The question matters because logistics decisions are rarely made by technology alone. They involve planners, managers, operations teams and, in many cases, customers or external partners. If these stakeholders cannot follow the reasoning behind an AI-supported recommendation, they are less likely to act on it.<\/p>\n
For Heinrichs, this could become a meaningful point of differentiation for European technology providers. \u201cWe can build AI as good as anyone, but we can add something different,\u201d he says. \u201cIt should not be a black box.\u201d<\/p>\n
As companies look to embed AI applications into established business processes, that difference becomes increasingly important. Systems need to be technically strong, but they also need to be understandable enough for users to challenge, validate and improve them over time.<\/p>\n
Managing Less Predictable Environments<\/p>\n
Operational environments are becoming harder to plan with historical data alone. Demand patterns shift, external factors intervene and market conditions can change quickly, often before those changes are clearly visible in the numbers. \u201cYou need to gather real-time data and not rely on historical data alone,\u201d he says. \u201cYou have to react to volatility and integrate signals from different sources into your decisions.\u201d<\/p>\n
This marks a shift from more static optimisation models towards responsive systems that continuously take new information into account. \u201cIt is getting more dynamic,\u201d he adds. \u201cThe next step is making it more agentic \u2013 reacting on its own to changes in the environment.\u201d<\/p>\n
From News to Forecast<\/p>\n
One example Inform presented for the first time at LogiMAT is a new AI-based approach designed to bring external events directly into forecasting and scenario planning. The starting point, Heinrichs says, was a simple question: why do forecast models so often ignore what is happening in the world around them?<\/p>\n
\u201cIf you run a classical forecast today, it is based on historical figures,\u201d he explains. \u201cBut in reality, demand is constantly influenced by events such as geopolitical conflicts, supply chain disruption, new regulation or market trends. This information exists, but usually as news, not as numbers.\u201d<\/p>\n
The new solution is designed to close that gap. Users provide a time series, such as sales figures or a market indicator, and briefly describe the context. The AI then researches relevant news events, analyses historical relationships and generates several possible future scenarios. The result is a forecast accompanied by an evidence-based explanation of why a market may develop in different directions.<\/p>\n
Human in the Loop<\/p>\n
![\"\"](\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/05\/26-IMAGE-1-1300x867.jpg\")
<\/p>\n![\"\"](\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/05\/GettyImages-1320009809.jpg\")
<\/p>\n