
Ignoring AI tools can have catastrophic consequences for a company’s efficiency and competitiveness, just as serious as hasty and forced implementations. Artificial intelligence is already at the top of investment priorities, yet a huge portion of projects still end in failure because companies lack strategy and clear goals, are not technically ready, buy fakes, or succumb to psychological market pressure. Little is also said about the fact that the technology itself is not perfect, yet very expensive, implementation is complex, and on top of that, an AI agent hallucinates and invents a reality that does not exist.
Robotic Process Automation (RPA), i.e., algorithms that perform repetitive rule-based tasks but lack the ability to learn. Then, advanced predictive analytics and machine learning (ML), and finally, generative artificial intelligence (GenAI), which exhibits a certain degree of creativity and can create new content based on simple language prompts. This is a simplified way to describe the increase in autonomy and the successive layers of intelligence added to various types of software. The last of these innovations proved to be quite a breakthrough, and according to IMF calculations, it is characterized by a historically unprecedented speed of adoption, reaching 100 million users in just a few months. This pace is significantly faster than that of other general-purpose technologies, such as the internet, cable television, or mobile telephony.
Today, everyone is talking about AI agents, a new approach to AI system architecture that combines individual layers and ranges of autonomy. From the information published by the creators of such solutions, one might even get the impression that an AI agent is an almost perfect, versatile, and indispensable tool, and that it is essentially harder to point out things an agent cannot do than those it performs with ease. The truth, however, turns out to be somewhat harsher than the marketing declarations of manufacturers.
AI agent, or what or who?
An AI agent is, in short, an autonomous system of algorithms using advanced machine learning and natural language processing techniques that allow it to understand the context of its environment, learn from previous interactions, select tools, and independently make decisions leading to the achievement of a set goal. All this can take place with human intervention kept to a minimum. Unlike commonly used process automation, which operates based on precisely defined rules, an AI agent, by analyzing patterns from the real world, many systems, and sources simultaneously, can adapt to them and take dynamic actions or issue appropriate recommendations. Whether an agent will have the ability to learn and modify its own behavior based on past experiences depends on its type, but there are already solutions that enable this.
AI agents can therefore learn the policy of their own functioning, e.g., based on a system of rewards or penalties received from the environment. The general principle, however, is that the agent receives a clearly defined goal, the realization of which it bases on data from the environment. It plans and executes this plan independently, and if necessary, performs operations in other systems and uses external tools, all the while striving to maximize efficiency. For a fuller understanding of this phenomenon, it is also worth adding that an AI agent is not a humanoid robot but software that is not meant to replace a human, but to enhance their actions in real time. Individual agents can work together, and humans supervising the work can always add context to the conclusions generated by such artificial intelligence systems.
The potential for using artificial intelligence was immediately recognized in supply chain operations
According to data published by Gartner, as many as 27% of supply chain leaders view investing in AI-based solutions as one of the three main factors for gaining a competitive advantage or eliminating a competitive disadvantage. For 9% of respondents, this is a top-priority goal. Similar conclusions are contained in an IDC report, according to which advanced analytics and AI in the supply chain were already the main investment priorities for the next three years in 2024.
What real and useful roles in logistics processes do AI agents offer today?
The key to understanding what an AI agent is and what tasks it can perform is to clearly distinguish its function from a chatbot or an assistant that passively reacts to commands. The agent acts independently; it must achieve the goal set by the human supervising it as quickly as possible. Only in this context can one see that the real potential of an intelligent algorithm, e.g., in contract logistics, lies primarily in areas requiring dynamic decision-making. A good example of a role for an agent is planning the work of people and equipment and continuously analyzing operational micro-decisions, such as allocating slots at ramps, balancing incoming and outgoing loads, or responding to unplanned shipments. A multi-agent system, representing different operational interests (e.g., Outbound, Inbound, Ramp, Guard, SLA), can efficiently work out an optimal decision using data processed by YMS, WMS, and TMS. This allows for shortening lead times, avoiding penalties, and improving indicators without engaging additional human or equipment resources. It should be emphasized that although this technology requires refinement, it is not just another trendy fad, but a transfer of real market mechanisms and predictions to a level that a human is unable to perform manually at such a volume.– explains Sławomir Rodak, R&D Director and Commercial Director at ID Logistics Polska, which provides comprehensive logistics and transport solutions, e-commerce support, and supply chain management in 19 countries.
It should also be noted that such implementations must be approached with proper preparation. An agent will not work effectively without a mature technological background in areas such as machine learning or computer vision. Solutions of this type will only make sense when they operate on reliable data and in an environment that enables them to make accurate and fast decisions. Without the right background, the agent itself will not change much in the organization; it may even generate more problems than benefits. Multiplication works both ways in this case. This is confirmed by numerous analyses in which the lack of a clear strategy, business goals, or proper input information is the cause of the termination of expensive implementations. At ID Logistics, we think of agents today as the next phase of developing our competencies; we focus on building the necessary technological foundations and developing projects that bring real business value. We are developing, among other things, automatic clothing dimensioning, employee demand prediction, and computer vision-based security systems. We are preparing our organization and technologies for the stage where agents will begin to realistically increase operational efficiency, not only as individual solutions but also as a network of cooperating systems.– adds Sławomir Rodak from ID Logistics.
The agent is capable, but needs a good background and equally capable people
The proactive capability of agents, which allows for faster and more accurate decision-making, has its limitations, and among them are those on the side of the organizations themselves. Data from PwC published in 2025 shows, for example, that for 37% of operations and supply chain managers, one of the three biggest challenges in effectively scaling artificial intelligence solutions is the availability and quality of input data. For 42%, it is the difficulty of integrating AI solutions with existing systems. This is not new, as similar problems could be observed when implementing other technologies in the supply chain, such as cloud solutions, which 56% of companies declare using, or digital twins, used by 21% of enterprises. Among those who attempted to implement digital solutions but ultimately declared at least partial dissatisfaction with the investments made, the main reason was complex integration (47%), followed by data problems (44%). Limited capabilities of solution providers (35%) and internal staff competencies (32%) also ranked high.
The dream of agents without one’s own data means a hard landing for ambitious assumptions
A number of entities, including those operating in the supply chain environment, are already presenting an exceptionally realistic approach to implementing agents. For example, FourKites argues that an AI agent without proper data will not be able to prepare clean and meaningful answers to the tasks entrusted to it. The difference between successful implementations and costly failures essentially comes down to data architecture and the work of agents as comprehensive control towers with real-time visibility into processes occurring in facilities, at suppliers, recipients, carriers, and customers. Without such an approach, this technology becomes just another warning system generating noise instead of value.
This logic is echoed by, among others, the World Economic Forum (WEF), pointing out that inconsistent, outdated, or unreliable data seriously limits the effectiveness of artificial intelligence, and older systems create additional integration obstacles. Regulatory requirements do not make things easier, so organizations must prioritize preparing their data for use in AI. Therefore, having a clear strategy for collecting external information that combines traditional operational data sources (inventory levels, tracking and location information), risk signals and supplier data (their health, geopolitical events, financial risk), and unstructured data (emails, meeting notes, market news) is of key importance.
The integrated set should create a comprehensive picture of the supply chain in real time. According to the WEF, such model platforms are still in their infancy, although this is a slightly brighter picture than the vision presented as late as 2024, when the organization suggested that it was very unlikely that a supply chain system would be fully automated in the near future. The main obstacle is the fact that it consists of dozens of industries and thousands of companies and is unable, or perhaps unwilling, to integrate its individual systems with the latest technologies. Full automation would also be slow and extremely capital-intensive, requiring cooperation that is difficult to achieve, especially in highly competitive areas such as maritime transport.
In this context, Gartner sees it a bit more optimistically, forecasting that by 2030, half of cross-sectional supply chain management solutions, i.e., those connecting multiple departments and functions, may use intelligent agents making independent decisions across the entire ecosystem.

Optimism is already spreading to senior management
A survey by the IBM Institute for Business Value (IBV) and Oxford Economics conducted among over 300 Chief Supply Chain Officers (CSCOs) and Chief Operating Officers (COOs) working in organizations implementing AI-based automation shows that they perceive AI agents as a business accelerator. As many as 62% of them realize that agents embedded in operational workflows accelerate action, decision-making, recommendation creation, and communication. As the analysis indicates, as many as 53% of executives are implementing work automation using AI agents in some form, 22% are developing conceptual versions, and 31% are already scaling them. Additionally, 70% of executives believe that by 2026, thanks to AI agents, their employees will be able to analyze data more accurately and support real-time optimization. More, 76%, claim that overall process efficiency will improve, and 57% expect AI agents to proactively create recommendations based on their knowledge.
Gartner cools these sentiments somewhat, arguing that while it is possible for up to 95% of data-driven decisions to be at least partially automated in such a short time, it still does not change the fact that only 10% of executives claim their companies use artificial intelligence strategically. Even fewer leaders (9%) declare that their organizations have a clearly defined vision for artificial intelligence, which is crucial for the smooth implementation, operation, and financing of solutions. Developing a comprehensive strategy for using AI in the supply chain in a breakthrough way, i.e., one that enables gaining a competitive advantage, includes four key pillars: vision, value, risk, and implementation. Only the assessment and implementation possibilities of these factors reveal the true potential of using the new technology, not only in supply chain operations but also in other areas.
What if AI agents are not as versatile and intelligent as their creators claim?
Data shows that sectors where the complexity of elements and their interdependence is relatively low benefit first from the young technology that AI agents are. According to 3,412 management representatives from various industries surveyed by Gartner last year, the greatest scope for using AI agents today is primarily in customer service departments (22%), with marketing departments (12%) far behind. Next are sales and business applications (11% each), as well as IT operations (10%). Only in 9th place are functions related to supply chains (5%), and even less probability of delegating competencies to intelligent agents is currently seen in the purchasing (3%) and legal (2%) departments.
A conservative approach prevails
Further analyses also indicate that not all managers have unanimously and widely opened their companies and wallets to implement AI agent solutions. Significant investments in this area are currently declared by 19% of respondents, and a conservative approach by 42% of executives. As many as 31% have opted for a wait-and-see strategy, and 8% have made no investments at all.
A wave of failed implementations is already sweeping through the market, and Gartner estimates that more than 40% of projects related to AI agents will be canceled by 2027, due to skyrocketing costs, unclear business value, and inadequate risk control. Additionally, a growing disparity between AI supply and demand is visible in the market, the so-called AI adoption gap. Solutions and innovations proposed by creators are growing significantly faster than the possibilities for their application on the client side.
According to analysts, this is just one of many factors that will lead to the bankruptcy of many intelligent algorithm providers in the short term, after which market consolidation and correction will follow. This will happen after the heavily exploited media hype subsides, and the psychological fear of missing out on new technology (FOMO) passes among potential recipients, and fundamental economic principles take over.
In the long term, specialized products regarding AI agents will emerge that will meet the real expectations of customers, and this is the normal life cycle of every product and technology. However, it cannot be ruled out that before this happens, AI agents will become another speculative bubble if investments detach from the technology’s potential to generate real and proportional economic value translating into concrete business results.
Fraud by the agentic FOMO method
The fear of missing out and implementation haste have already led to serious pathologies, which in the future may discourage companies from attempting implementation. There is a FOMO phenomenon in the market, but also so-called agentic washing, i.e., a mechanism of falsification in which software vendors present existing products as AI agents. These are, for example, masked virtual assistants, advanced RPA systems, and chatbots. Gartner estimates that globally, only 130 creators out of literally thousands of software vendors advertised as AI agents are genuine. Most applications pretending to be agents lack business value or ROI because the models have not yet entered the maturity phase and the stage of truly autonomously realizing complex business goals. In a word, the market has been flooded with fakes.
The agentic bubble has its problems, but if it doesn’t burst, it won’t stop
According to Gartner, in 2028, after consolidation and reaching greater maturity, at least 15% of daily business decisions will be made by autonomous agents, which will be a significant increase compared to 2024, when this share was 0%. Additionally, 33% of business software applications will contain AI agent algorithms, which will be a jump from less than 1% in 2024. For now, however, technology sector analysts recommend that at this stage of development, the implementation of agent-based artificial intelligence should be started only when it brings clear value and return on investment. It should also be remembered that the integration itself with existing systems is technically complex, expensive, and often disrupts the company’s work.
The maturation period for agents will be very dynamic, but at the same time very profitable. In the best-case scenario, it is predicted that by 2035, artificial intelligence based on them will generate about 30%, or over $450 billion, of the revenues of companies producing business applications. This is significantly more than in 2025, when this share was 2%. New functionalities are also expected to become widespread every year. In 2027, individual specialized agents will cooperate with each other to solve complex tasks, and this will be one-third of implementations. A year later, they will create ecosystem networks and react dynamically in a changing environment. In 2029, we will enter a new dimension of cooperation between agents and humans, and at least 50% of knowledge workers will develop new skills needed to work with, manage, and even create agents to solve complex tasks.
Before this futuristic scenario is realized, the AI agent must stop making things up
Yes, an AI agent makes things up, or rather hallucinates, because that is the technical term for generated results that have little to do with reality. Hallucinations are one of the more serious factors hindering the development of AI; they are widespread and, for now, difficult to eliminate. Google points out, for example, that incorrect results can be caused by insufficient training data, their bias, prejudice, or incorrect assumptions of the AI model. It may simply not understand information about the real world, physical properties, or certain facts, and if it learns incorrect patterns during training, it will lead to hallucinations in the form of results that are seemingly credible but in reality untrue or nonsensical. OpenAI explains that models hallucinate because standard training and evaluation procedures reward guessing instead of admitting ignorance. IBM, in turn, explains that model hallucinations are similar to how humans recognize silhouettes or shapes in clouds, in a way adding context to what they see.
So far, it has not been possible to eliminate the problem of AI model hallucination, but its scale is truly serious. This leads not only to the generation of false results but also undermines trust in solution providers. According to IT department heads of companies employing at least 250 people surveyed in mid-2025 by Gartner, this is one of the main barriers preventing the implementation of fully autonomous agents. Only 19% of respondents showed high or complete trust in their provider’s ability to ensure protection against hallucinations. Significantly more, 26%, however, believed that agents would have a transformative impact on productivity, and 53% believed that their impact would be at least significant.








