Artificial Intelligence and the Transformation of Modern Supply Chains
- ISCRO MSU
- 20 hours ago
- 13 min read
Written by Aryahi Pachpande (Research Lead), Arshdeep Singh, Ajitesh Venkatesh, Arham Jhaveri
EXECUTIVE SUMMARY:
AI adoption across U.S. businesses has steadily increased, rising from roughly 2.5–4% in early 2023 to 5–9% by late 2025, reflecting a broad shift toward machine-learning forecasting, demand sensing, warehouse automation, and predictive planning
Automation, robotics, and advanced planning tools are transforming warehouse and production operations, while AI also introduces new cybersecurity vulnerabilities as attackers use automation, deception, and adaptive malware to exploit supply chain networks
Growing AI adoption also expands cybersecurity risks, as attackers use AI to scale threats, exploit supplier weaknesses, and create highly targeted, evasive attacks
AI is becoming a major driver of productivity and sustainability, helping firms reduce waste and emissions, though challenges like data quality, energy use, and implementation costs remain
Introduction
In late 2022 and early 2023, the rapid rise of artificial intelligence introduced a major turning point for global industries, including supply chain management. Recent research estimates that generative AI could generate between US$ 2.6 and 4.4 trillion in global economic value each year, underlining the technology’s far-reaching impact (McKinsey). Tools like ChatGPT and advanced machine-learning models made AI more accessible than ever before, pushing companies to rethink how they forecast demand, plan operations, and respond to disruptions.
This shift arrived at a critical moment, as supply chains were still recovering from the impacts of Covid-19, extreme weather events, and growing geopolitical tensions. Traditional forecasting systems, which relied heavily on historical data and steady market conditions, can struggle to adjust to fast-changing and unpredictable environments.
From 2023 to 2025, AI adoption expanded quickly across major industries, with businesses using it to interpret real-time signals, predict demand more accurately, automate planning, and build resilience. This marks the beginning of a larger transformation in supply chain management, with AI becoming a core part of how companies plan and operate.
AI in Forecasting
Currently, many supply face greater levels of unpredictability from various parts of firms’ external environments (e.g., regulatory changes, rapid swings in consumer demand) that have exposed the weaknesses of traditional forecasting models. The disruptions have made it clear that companies can no longer rely on just human judgement or outdated forecasting systems to make predictions. These traditional systems were built for more stable environments and are heavily dependent on extracting information from historical data, which is often less effective in a evolving market. Due to this, they struggle to adapt and predict signals such as social trends, consumer behavior, and weather patterns.
The pressure to improve AI forecasting is a growing trend across all industries, with IBM claiming improving forecast accuracy is now one of the top three priorities for CEOs over the next three years (IBM). It is clear as the market continues to rapidly change; new forecasting models will be essential for success. Additionally, 88% of retail executives believe demand forecasting is the area where AI can have the most dramatic impact (IBM). This pressure has pushed companies to implement and utilize AI models for greater accuracy and adaptability.
This chart illustrates the steady rise of AI adoption in U.S. businesses over the last two years, based on BTOS industry survey data (U.S. Census Bureau). Even though monthly fluctuations occur, the overall trend shows clear growth in both current AI use and expected AI adoption in the near future. The upward trend in expected future AI use reflects how organizations are preparing to rely more heavily on machine-learning-driven prediction, demand sensing, and real-time planning to manage supply disruptions (Figure #1).
AI can enhance forecasting accuracy and efficiency through its various advanced machine learning capabilities. Unlike traditional models, AI models can identify and adapt to changing conditions, reducing forecasting errors, with IBM reporting that AI has been shown to cut forecasting errors by up to 50%. This derives from AI’s major strength in its ability to consolidate data from a wide range of sources including manufacturing, operations, supply chain systems, and other external factors. For example, Walmart utilizes AI demand-sensing to merge weather data, local events, and purchasing trends to improve predictions at the store level (IBM).
AI provides improvements to operational efficiency, through reducing the time and labor required to produce forecasts. For instance, The Idaho Forest Group was able to reduce its forecasting process from more than 80 hours to less than 15 hours by implementing AI, demonstrating how automation significantly speeds up planning cycles. Similarly, global agribusiness manufacturing implemented c3.ai, seeing schedule generation times fall by 96% and saving over $1.5 million through fewer production changeovers. Implementing AI not only improves accuracy but also frees employee time through automation and cuts down costs.
Despite these various benefits of implementing AI, companies still face several challenges and limitations. With data quality remaining as a major concern, as AI systems depend on accurate, complete data, with IBM noting that “inaccurate or incomplete historical sales data can undermine even the most advanced AI models” (IBM). AI models rely heavily on accurate data to be successful, if lacking this information AI systems will struggle to give effective results. Additionally, implementing AI models requires the unification and integration of several different data sources. C3.ai’s agribusiness client, for example, had to unify 18 separate data sources before AI forecasting became fully functional. This process not only requires substantial financial investment but also technological expertise (C3.ai).
Looking into the future, AI is expected to have a significant impact and forecasting and demand planning within the supply chain space. Larger corporations that can afford to take on investments and challenges to implement these new advanced models will gain a competitive advantage over smaller corporations. As we see AI become common within the demand forecasting space it will become commonplace and will go beyond forecasting as these tools continue to advance into other sectors such as automated production scheduling, creating increasingly integrated and intelligent planning ecosystems.
Automation and Robotics
AI agents now handle real-time inventory tracking, coordinate restocks, manage material selection, and safely transport heavy materials, while robots support picking, transport, and shelf-movement activities (Figure 2). This reflects how modern warehouses blend human labor with robotics and intelligent systems to increase speed, accuracy, and productivity. As companies face rising labor shortages and pressure for faster fulfillment, automating manual processes has become essential for staying competitive. While high upfront costs and concerns about job displacement make some firms hesitant, failing to automate can be just as costly. Understanding the tradeoffs reveals that automation is not just a movement or trend, but a strategic move to boost profits (Raymond).
As companies adapt to other changes, robotics (stock IQ) has become increasingly prevalent in the manufacturing industry. Furthermore, robot installations are set to increase by 50% in large-area warehouses. It is also projected that the overall robotics market will grow, changing the supply chain's market share in the sector. Not only are robots taking over supply chain operations, but the market is projected to reach 41.7B by 2032. The sudden surge also affects the labor market, which, in essence, raises operational demands and takes a big leap toward what else robots can do in the near future. It is also reported that, as a result, 37% of companies have proposed layoffs and the replacement of workers with robots to fill blue-collar gaps. In the market, there are many types of supply chain automation machines, for example, the AMR (Autonomous mobile robots) specialize in moving with sensors, and Cobots (collaborative robots) work simultaneously with humans. Both types of autonomous machinery are set to grow steadily in the coming years. With such robots entering the global market, there are also costs associated. Sources estimate that fully automated facilities can cost upwards of $25 million, but the range is $5-10 million based on value and operational use.
This chart shows how different U.S. skill categories vary in their automation exposure, and it highlights why warehouse and manufacturing roles are being transformed so quickly. Skills tied to physical and routine tasks face the highest exposure, with categories like handling and moving showing 42% moderate and 29% high automation exposure, and machinery-related work showing 55% moderate and 27% high exposure. Constructing work also shows elevated exposure, with 44% moderate and 31% high (Figure #3). Overall, the numbers show that physical and repetitive tasks are the most likely to be automated, while higher-skill cognitive areas will shift toward AI-supported roles rather than full replacement (McKinsey).
Cybersecurity
The integration of Artificial Intelligence to enhance supply chain processes is revealing a major security gap as businesses tend to overlook the risks that come along with integration. AI is a "dual use" technology. On one hand, it brings benefits to businesses, on the other hand it equips attackers with powerful tools, changing the risk landscape fundamentally (IBM). It's important to recognize that this risk is not just hypothetical. It carries the very real potential for major financial setbacks and regulatory action. For instance, the EU AI Act would penalize as much as €35 million or 7% of the global revenue for serious security failures related to AI. This law highlights the issue of AI security as it presents more threats than opportunities across the supply chain (InterosAI).
AI-powered attacks deliver incredible results because they focus on the weakest links in ecosystems that large companies count on thousands of small vendors with lower security maturity. To speed up their efforts and scale, attackers employ AI for automated reconnaissance, as it’s shown from a 156% rise in the number of malicious package uploads to open-source repositories during one recent year. This explains AI's role in automating the building of malware that cannot be detected by conventional methods. The 2025 "SolarTrade" logistics breach is a case where reports suggest attackers used AI to insert malware code in a routine software update thereby crippling the operations for several months (InterosAI).
A detailed attack on the Salesloft Drift program shows how attackers take advantage of third-party trust and authentication (UpGuard). This event, blamed on a group called UNC6395, was a textbook supply chain assault that resulted in the theft of safe “OAuth tokens”. These tokens represent "the new keys to the kingdom", and they set up long-term rights over cloud applications. The intruders, by obtaining access to Salesloft's GitHub account, stole these tokens and, and enjoyed the same admin-level access in customer ecosystems from companies like Salesforce and Google Workspace. The eventuality serves to show how a single vulnerability in a vendor's environment can become an opening for an attack that leads to another company's trusted network.
The actual purpose of the Salesloft intruders was to retrieve sensitive data concealed in the most unsuspecting places, like customer support cases, written in plain text. It reveals a crucial defect in the defensive front as hackers are aware that such simple text covers an abundant source of high-profile credentials. The perpetrators went on to take API keys, cloud credentials, passwords through illegitimate channels, all the while they kept removing query logs so as not to trigger alarms. This attack is made easier by the fact that the average time for breaches to be discovered is 276 days, during which the breaches have already existed. The time is mostly extended due to stealthy AI-assisted attacks (InterosAI). This single breach compromised hundreds of organizations, among which Palo Alto Networks and major tech firms like Cloudflare are included (UpGuard).
Besides improving existing techniques, AI brings in completely new attack angles that are aimed at the models themselves. The main vulnerabilities are associated with three stages: Input (Data Poisoning), Model (Algorithm Corruption), and Output (Prompt Injection). From the stages listed, data poisoning tricks the AI to accept falsified data as truth for it to make decisions on that basis, whereas prompt injection tricks an LLM into giving a response that contains protected information or performs actions without permission. Present-day security infrastructures are often inadequate in addressing such AI-specific challenges as covert neural backdoors (IBM). Although AI can be a very powerful offensive weapon, it can be used defensively as well to improve threat intelligence and provide better visibility over supplier risk thereby contributing to overall operational resilience.
Macroeconomic Effects of AI Adoption
AI has begun to have a measurable impact at the macroeconomic level, serving as a major driver of productivity across global industries. McKinsey estimates that AI has the potential to automate 60%-70% of tasks that currently require employee time, boosting productivity, and allowing employees to invest their time in other, potentially more complex tasks (McKinsey). Even with AI’s relatively recent rise, it is estimated that around 40% of the world’s current GDP is connected to activities exposed to AI automation (UPENN Wharton). These continued rising trends show how integrated AI has become at the global level and how it continues to shape the economy.
Estimates show AI’s quantitative impact on the global economy. Top firms like McKinsey estimate that generative AI creates $2.6–$4.4 trillion in economic value annually across major industries, with sectors such as banking potentially adding $200–$340 billion annually through the implementation of AI systems. Additionally, based on current capabilities, roughly 23% of tasks in AI-exposed sectors can be fully automated, boosting productivity and having long-term economic benefits (McKinsey). These estimates reveal that AI’s impact, when applied across different industries, will significantly influence productivity and the global economy, leading to meaningful improvements in GDP over time.
Despite AI’s limitless potential, there are challenges and limitations that shape how quickly AI’s benefits will be realized. AI is rapidly changing and evolving, continuing to outpace the current workforce's skills, leaving many users behind and creating gaps that limit effective adoption within the workplace. Due to this it is unclear if current estimates accurately predict AI’s impacts, with the Penn Wharton Budget Model noting that economic forecasts that involve AI are highly sensitive to assumptions, as the technology is evolving too quickly to create precise models, creating ambiguity on whether AI will be as significant as estimates predict UPENN Wharton). Additionally, AI’s gains across all industries will not be evenly distributed, with industries like tech or banking being more aligned to AI tools compared to labor-intensive industries. Leaving some industries economically behind, while others advance quickly with AI development.
Even with these challenges and limitations, estimates predict that AI is expected to provide steady growth to productivity that will compound over the next few decades. With AI automating repetitive tasks, workers will be able to divert their time from simple and repetitive tasks to work on more complex, creative tasks. The short-term returns of AI may be uncertain, but the long-term AI's impact on productivity and GDP will be significant. AI has the potential to become one of the primary drivers of future economic growth.
Sustainability
Sustainability has significantly influenced the procurement and logistics of large companies. Corporate businesses have shifted their focus to become more environmentally friendly to address growing environmental concerns. This transformation is reshaping how industries operate. Consumers demand a shift from old practices that harm the environment, and companies must move to mitigate costs, use ethical sourcing, and reduce their overall carbon footprint. As innovation and competitiveness have increased, sustainability has become a main focus point, and the use of artificial intelligence has helped tremendously in this effort.
Relex is a tool used by grocery retailers to reduce waste and optimize inventory in dynamic supply chains. Origin, a Spanish-based grocery retailer, oversaw a 10% increase in product availability and a 30% reduction in food spoilage. Migros Online, a fully online grocery platform, experienced a 20% reduction due to services provided by Relex (Relex).
With the rise of AI in sustainability, many challenges are also emerging. Although using AI in the sustainability market is considered very efficient, power consumption is a major concern. Since the tools used by major retailers are often extensive, their use can lead to overconsumption. It is reported that AI can consume up to 33 times more energy than a regular processing unit. Another issue arising from AI in sustainability is higher power consumption (AB Magazine). Because AI must forecast and handle a large variety of free-flowing data related to crops or materials, this data may not be stored properly, leading to disorganized collection and reuse for future purposes. Additionally, a challenge with sustainability of AI agents is the lack of scalability. Scaling AI solutions can be very difficult, as the same solutions often do not work for every problem. Even if AI adapts well, smaller companies may struggle to implement the solutions needed to improve sustainability.
Overall, this sustainability sector, not just in the supply chain but across various industries, has grown rapidly. With the use of AI agents, carbon emissions and regulatory demands are being addressed. While generally, generative AI appears to be efficient and is encouraged by our government, there are still regulations and laws that companies must follow to avoid excessive energy use (Deloitte). These factors not only influence companies but also affect buyers because when firms need to comply with regulatory demands, the shift often impacts consumers, who are frequently burdened with tariffs or higher prices for goods and services.
Conclusion
From 2022 to 2025, AI has rapidly shifted from a new idea to a fundamental part of modern supply chains. Adoption in U.S. businesses rose from roughly 2.5-4% in early 2023 to 5-9% by late 2025, showing how quickly companies have embraced AI tools to strengthen forecasting, automate planning, and build resilience. These systems now help firms interpret real-time signals, adjust to market volatility, and reduce forecasting errors by up to 50%. Companies like Walmart and Idaho Forest Group demonstrate how AI-based demand sensing and automated planning can merge weather data, local events, and operations data to cut planning time from over 80 hours to less than 15.
Across different areas of the supply chain, AI continues to reshape how companies operate. Forecasting tools help firms interpret real-time signals and adjust to sudden market changes, while robotics and automation are improving warehouse efficiency and reducing manual work. At the same time, AI has created new cybersecurity risks, as attackers use advanced automation, deception, and adaptive malware to exploit gaps in supplier networks. These challenges show that while AI brings major benefits, it also requires stronger data governance and more secure systems.
Beyond operations, AI is influencing larger economic trends and sustainability efforts. It is pushing productivity higher, helping companies reduce waste, and supporting more environmentally conscious planning. But these gains come with limitations, including data quality issues, energy demands, and uneven adoption across industries. Overall, the rise of AI marks a meaningful turning point for supply chain management, and its role will only continue to grow as businesses work to balance its opportunities with its challenges.