Log-hub chief sees AI widening supply chain analysis

Log-hub Chief Executive Officer Jan Sigmund has outlined his views on the use of artificial intelligence in supply chain analytics during an appearance on the Supply Chain Connect podcast. He said the technology is changing how planners interact with analytical tools.

Sigmund described a shift away from specialist software skills and towards framing business questions more clearly, as AI tools take on more of the work of interpreting requests, cleaning data and presenting results in plain language.

The discussion reflects a broader debate across business software about whether generative AI will replace specialist roles or change how they are carried out. In Sigmund's view, supply chain planning is likely to follow a pattern already seen in software development, where work moves from direct production towards supervision and direction.

He compared the shift to coding, arguing that programmers have not disappeared even as AI tools have taken on a larger role in writing software. Instead, their work has moved towards guiding systems and checking outputs, and he expects a similar change in supply chain functions.

That shift matters in a field where advanced analytics has often been limited to specialist teams, external advisers or companies with large budgets. Network design and other planning exercises have historically required expensive licences, technical expertise and long project timelines, making them difficult for many companies to run in-house.

Sigmund argued that easier interfaces could widen access to those tools. He linked that idea to Log-hub's products, including supply chain applications built as an Excel add-in, as well as the company's integrations with Claude and ChatGPT.

Access shift

His underlying argument is that AI may lower the barrier to entry for staff who understand the business problem but do not have deep training in specialist modelling software. Rather than learning a complex interface or scripting environment, users may increasingly describe the issue they want to solve and then evaluate the system's suggestions.

"Democratizing is not dumping down, it's a better packaging of the solutions," said Jan Sigmund, Chief Executive Officer, Log-hub.

"When analysis gets cheaper, people ask more questions," Sigmund said.

This view emphasises volume and accessibility rather than replacing existing analytical methods. If the cost and time needed to run a scenario fall, companies could test more options in areas such as network design, sourcing, inventory or transport planning, although Sigmund argued that decision-making should remain with people.

Human oversight

On that point, he drew a distinction between AI's role in handling language and data preparation and the continued use of established optimisation engines for the mathematical core of planning work. He argued that large language models are suited to understanding questions, cleaning messy inputs and explaining outputs, but should not be treated as the final decision-maker.

He compared the role of AI to a car navigation system. In that framing, AI proposes a route and explains the logic, while the person remains responsible for deciding whether to follow it.

That position mirrors a common stance among enterprise software providers seeking to incorporate generative AI without surrendering control of business-critical decisions to probabilistic systems. In supply chains, where errors can affect stock levels, transport costs and service performance, the line between recommendation and action remains commercially significant.

Log-hub, founded in Switzerland in 2017, works with more than 180 companies worldwide and has a team of 50 across Switzerland, Germany, the US, Serbia and India. Its supply chain applications have been downloaded more than 30,000 times.

Its focus on packaging analytical tools for wider use also reflects a longer-running industry trend. For years, software suppliers have tried to move advanced planning and modelling out of the hands of a relatively small number of experts and into day-to-day business teams. Generative AI offers a new route towards that goal by using natural language as the interface.

Whether that translates into sustained adoption will depend on reliability, data quality and companies' willingness to let staff use AI systems in operational settings. It will also depend on whether suppliers can show that these systems improve access without weakening the quality of the underlying analysis.

For now, Sigmund's argument is that the value lies not in handing over judgement to AI, but in reducing the friction involved in asking better questions and getting understandable answers from existing analytical systems.