Skip navigation

M&M: Using artificial intelligence to clear resources

|   M&M

Innovative technologies play a decisive role for a modern service provider like Militzer & Münch. At the moment, the use of artificial intelligence (AI) is an issue that is causing a stir in the logistics industry. In order to explore the individual needs and potential applications of this technology, Thies Spannagel has been appointed Group Project Manager AI. His task is to evaluate processes where AI can be used to reduce the workload on employees and increase the service level and efficiency of the logistics service provider. Thies Spannagel has been with Militzer & Münch for over six years and has more than 20 years of professional experience in the logistics industry.

Mr. Spannagel, what is your task in the AI project?
As Group Project Manager AI, I am like a hub. I receive input from the country units telling me what challenges they have to overcome on a daily basis, and where major expenses arise that could be reduced using AI. The focus is always on improving the quality of service for our customers. I then check whether there are any suitable AI modules that we can integrate into our IT infrastructure. I’m very interested in AI because I’m very process-oriented and want to achieve my goals fast and efficiently. If we process the orders in a clean and orderly manner, we achieve a high level of customer satisfaction.

When was the project launched, and what is the objective?
We started the project on 1 March. Of course, I had already been looking into the topic of AI before then. In November 2023, M&M air sea cargo GmbH took a new transport management system into operation, our central software that we use to process all orders. The task now is to further develop this system – including with AI. The aim is to continuously reduce for our colleagues the number of manual data entries in order to increase productivity. However, we are also looking at other areas where AI can support us.

Where do you see the biggest challenges?
The Militzer & Münch Group’s country units are relatively heterogeneous. Each country has its own transport management system and its own IT infrastructure. This is a challenge when we purchase new modules, as we have to ensure that they can be used by as many country units as possible. Evaluating the modules is also demanding. Service providers often present demo versions in their showcases. The scenarios shown in these demos are often perfectly tailored to the tool’s capabilities. However, the problems we face in reality often have completely different parameters, for which the AI must first be trained over a lengthy period of time before we can use it – a new AI module is never ‘plug and play’.

A good example is something we saw with one of our partners in Canada. They wanted to implement a module that uses AI to process customer enquiries quickly and efficiently. When I looked at this with my colleagues, I was initially really impressed. After a few months, when we asked if our partner was happy with the tool, they said it wasn’t being actively used yet. There were still too many early problems and the tool was not even suitable for many areas in the company. Manual intervention was required so often that it was more efficient for employees to process enquiries manually.

When we select providers, they often ask in advance for examples from our company that they can use to demonstrate their tool. However, we also always bring along a few cases that the provider is not yet familiar with. That way we can better assess whether we can actually use the tool or whether the provider has only customised the demo to our specific case. Risk management is also my task when selecting these tools. That’s why for such meetings I take up to three experts with me who deal with the issues under review on a daily basis, so that we don’t conclude contracts lightly.

What characteristics do AI modules need to have in order to be considered for Militzer & Münch?
At the moment, we are primarily interested in using AI for repetitive tasks in order to free up resources in other areas. We can thus optimize processes and become more efficient.

A good tool has to fit in well with our IT infrastructure, which is of course difficult due to the different systems we have in the individual countries. When I talk to the provider of an AI module, I naturally always ask what interfaces are available. The API standard, which has the advantage that the code does not have to be rewritten from scratch during implementation, is already widely used. We have to check which providers are available on the market and for which areas they are suitable. One example is customs clearance. The aim is to automate the process so that the documents we receive from the customer are already prepared to such an extent that our colleagues only have to look over them again and add small details – but don’t have to search for and enter all the information themselves.

With technologies such as AI, it’s important to be involved right from the start and follow the development of promising modules so as not to be left behind later on. My job is to play a leading role in supporting my colleagues in the other countries in implementing the modules. However, the respective country unit must provide sufficient resources – in particular a temporary implementation team of local experts who are familiar with the regulations of the respective country. Teamwork is absolutely called for here.

AI systems rely on data, and the data must be prepared in such a way that the AI can process it. Last year, for example, the colleagues at the German Road (M&M Militzer & Münch GmbH) began using an AI module that extracts orders from major customers’ emails and automatically creates an order in our transport management system, largely eliminating the need for manual input. However, it took a year to train the system so that it works for us, which tied up a lot of resources.

Which institutions or service providers are supporting Militzer & Münch in the project?
We cooperate with the University of St. Gallen. In this context, we also exchange ideas with a student consultancy that also collaborates with the university. The consultancy has already supported AI projects. Their experience is very valuable to me, because we don’t want to make mistakes that we only notice years later. There is a lot of potential for mistakes, especially in the field of innovative technologies. That’s why it’s important to get input from different areas. We need to find out which technologies are on the market, which areas they cover and what potential they have. The price-performance ratio and the options for implementation in our IT infrastructure also play an important role. Only once we have gathered information about all of this can we decide whether an AI tool or module is of interest to us.