Digitalization in logistics is extremely uneven: the most advanced companies are testing complex technologies, but most companies are at the initial stage of automating simple operations. For example, half of enterprises have not yet switched to electronic document management and do not use IT platforms to search for cargo.
At the same time, artificial intelligence (AI) is developing rapidly – and this technology can no longer be ignored. Thus, the volume of the global AI market in logistics in 2024 amounted to almost $18 billion, and in 2025, it will grow to $26 billion. In the next 10 years, the market will grow by an average of 44.4% per year.
Russian companies are eyeing technology and making plans. According to a study by Strategy Partners, 45% of logistics market participants intend to implement AI in the next 2-3 years. We are talking about planning processes, cargo delivery and customer service.
Key AI technologies in logistics
Artificial intelligence is a broad concept. Three AI technologies that are most in demand in logistics can be distinguished.
Machine learning (ML) is used to work with data. These solutions help predict demand and prices, optimize inventory, and manage risk in complex supply chains. Algorithms take into account historical data and macroeconomic indicators, allowing to simulate various scenarios for the development of events. In addition, ML-based systems help maintain equipment. For example, FedEx’s predictive maintenance platform analyzes data from more than 35,000 vehicles, reducing fleet maintenance costs by $11 million per year and reducing downtime by 22%.
Computer vision is used to automate goods receipt, accurately record cargo, and monitor safety standards. Such systems recognize damage to packaging, verify labeling and control the storage of goods. This minimizes human error. In addition, computer vision is used in autonomous systems – from warehouse robots to unmanned trucks. In Amazon’s order centers, for example, computer vision systems control robots that transport shelves of goods to human pickers.
Generative AI is the 2025 trend. The volume of the global GenAI market in logistics has already reached $1.36 billion and will grow by 36% annually until 2035. Its main function is the autonomous execution of operations. For example, workflow: a neural network classifies invoices, checks declarations and identifies inconsistencies, speeding up cargo processing. Smart chat bots process up to 80% of calls without human intervention, instantly answering questions about the status of delivery, the availability of goods and the optimal route.
AI in practice
Several scenarios for the use of AI in Russian companies can be distinguished – these are the solutions that will be actively tested in the coming years.
1. Autonomous trucks
90 unmanned trucks of the third level of automation are already running along the country’s highways (the driver is in the cab). On the Neva M-11 highway (Moscow – St. Petersburg) in 2025 alone, unmanned vehicles made 21 thousand trips.
Safety and efficacy indicators are encouraging. Trucks traveled more than 8.7 million km along unmanned corridors on the M-11 and Central Ring Road without an accident.
Autonomous tractor Navio overcame the path from St. Petersburg to Kazan with a length of 1600 km in 24 hours. This is almost 2.5 times faster than a standard flight with a driver, who must comply with the work and rest regime and spends an average of 58 hours on the route. Navio reports that it has transported more than 130 thousand tons of cargo in this way, while reducing fuel consumption by 15%.
2. Customer service
Generative AI is fundamentally changing the approach to customer service. Modern chatbots are capable of dialogue, coping well with typical questions about the location of cargo and the state of financial settlements.
When non-standard situations arise, the AI seamlessly transfers the client to a specialist. This reduces the burden on call center employees and speeds up the consideration of applications many times over.
In world practice, by the end of 2025, chatbots independently process up to 80% of routine requests – tracking cargo, checking the status of payments and consulting on tariffs.
3. AI dispatchers
AI controllers provide comprehensive real-time transportation monitoring. Modern systems track not only the location of vehicles, but also technical parameters: from fuel consumption to temperature in refrigerators. This saves resources significantly. All data is integrated into a single transport management system, creating a transparent analytical base for optimizing routes and predicting potential failures. At the same time, AI does not replace a human person, but strengthens his capabilities, taking on routine tasks and leaving to specialists to solve complex and non-standard situations.
4. Voice control
Voice interfaces will replace manual data entry on freight exchanges. Instead of filling out complex forms, users will be able to formulate requests verbally like “Show free trucks from Krasnodar to Moscow for Wednesday.”
Natural language processing technologies provide instant speech recognition, accurate parameter extraction, and instant data entry. This will speed up the search for vehicles and ordering several times.
5. Forecasting
AI-based decision support systems analyze heterogeneous data in real time, from transport IoT sensor metrics to macroeconomic indicators and weather data. They not just predict demand, but also model various scenarios for the development of events.
For example, for perishable goods, the algorithm analyzes the current state of the goods, route congestion and market conditions, suggesting the optimal solution – to continue delivery, redirect to a warehouse or issue a return.
Pragmatic approach to implementation
When planning the introduction of AI into logistics processes, it is worth starting with small but indicative projects – those where you can quickly achieve business effects. This can be a chatbot for processing requests, a demand forecasting system for one product category, or a pilot project to optimize routes for part of the fleet.
An important principle of successful implementation is the passed stage of basic process automation. If the electronic document flow is not launched, the automatic search for transport and cargo is not mastered, there is no real-time transportation tracking system – you should start with these tasks. And only after that, plan the introduction of AI.
At the same time, it is important to understand: it does not always make sense to develop a complex system yourself or buy an expensive solution. Small and medium-sized businesses have enough AI technologies already built into logistics platforms. In most cases, it is easier to use affordable solutions using the SaaS model (software as a service), which do not require long implementations and large investments.
By Svyatoslav Wilde, founder and director of the ATI.SU Freight Exchange