- Will AI be able to replace freight forwarders?
- What skills should be developed?
- How can we benefit from this new technology?
Artificial Intelligence (AI) has been developing at a rapid pace in recent years, offering surprisingly powerful capabilities in many fields. The TSL (Transport - Freight Forwarding - Logistics) industry, due to its complexity, is a challenging area for AI applications. As a result, compared to other industries, breakthrough systems using the latest technology have not yet become widespread in this field. However, it's only a matter of time.
The national exam for freight forwarders requires candidates to possess extensive knowledge of transport procedures, international law, supply chain management, and analytical skills. The written part of the "AU.31: Transport Organization and Customer and Contractor Service" exam consists of 40 tasks, with 60 minutes to complete them. Sample questions from 2024:
Thinking has a future
Just a year ago, AI had difficulty passing this exam. However, in early January 2025, there was a breakthrough. This is thanks to new techniques that allow more complex problems to be broken down into smaller steps. While AI still operates on text patterns, it can now create a series of reasoning, which gives the impression that it is "thinking." This enables programmers to program AI by indicating a methodical approach to solving specific tasks. This, in turn, has led to even better results.
A research project carried out by RST Software in collaboration with Poznań School of Logistics achieved an average score of 84.3%, which means passing the professional freight forwarding exam.
The AI system created uses a model trained on vast datasets, including those related to transport and forwarding. Thanks to machine learning algorithms, AI can recognize patterns, analyze text, tables, and images. During the exam, the system had to solve cases simulating scenarios from the freight forwarding industry, such as route planning, inventory management, and resolving logistics issues.
The system was tested on 1,550 questions from exams between 2016 and 2024. It passed every one of these exams! The best result—89.4%—was achieved on the 2019 exam.
if you'd like to know how this study was conducted, you can read our article that outlines the technical aspects of the experiment and explains how we trained the model.
Will AI be able to replace freight forwarders?
Undoubtedly, the capabilities of this technology grow every month. Even today, it can read documents, calculate costs, and input data into various systems faster and better than humans. It speaks and understands all languages and can listen and speak in a manner indistinguishable from humans for the past two years.
Therefore, it is ideally suited for automating tedious tasks or processes that require quick action, such as:
- Managing order documentation, invoices, and waybills
- Vehicle planning and route optimization
- Automatic order monitoring and status updating
However, AI still struggles in areas requiring intense human interaction and deep understanding of emotional or cultural context. This includes negotiating prices and terms, building relationships with contractors, resolving conflicts, and reacting to non-standard situations in real time.
In the TSL industry, not only price and deadlines matter, but also reputation and the reliability of on-time, safe deliveries, which is often achieved through long-term, personal relationships with partners.
What AI skills should be developed?
AI will quickly take over more and more operations, increasing the efficiency of freight forwarders' work. This will intensify competition, requiring openness to new technologies and continuous improvement in using the latest tools.
While AI will not replace a competent freight forwarder, it can be a great support in data analysis and automating repetitive tasks. Understanding new tools and being ready to implement them will help stay competitive in the market. AI will not replace freight forwarders – other freight forwarders who are skilled in using it will.
.png)
