MIT CTL–Mecalux study finds rising adoption of generative and agentic AI in warehouse operations, with 90% of companies planning to increase AI investment over the next two to three years.
A new global study by the Massachusetts Institute of Technology (MIT) Center for Transportation & Logistics (CTL) and Mecalux highlights the growing role of artificial intelligence (AI) in warehouse operations. According to the report, 57% of organisations expect AI to make or execute more than a quarter of their operational warehouse decisions autonomously by 2028, while 90% plan to increase their AI investments over the next two to three years.
Titled The State of Generative and Agentic AI in Warehouse Operations, the study was conducted by MIT CTL’s Intelligent Logistics Systems Lab in collaboration with Mecalux. It draws on responses from nearly 2,000 supply chain and warehousing professionals across 21 countries, examining AI adoption, operational performance, decision-making autonomy and workforce developments.
AI Adoption Expands Across Warehouse Operations
The survey indicates that AI adoption is already widespread, with 61% of organisations reporting large-scale deployment or full integration of generative AI (GenAI). For agentic AI, the corresponding figure stands at 56%.
However, most implementations remain relatively recent, with approximately three in four applications having been at their current deployment stage for less than a year.
The reported operational benefits are encouraging. Around 87% of GenAI use cases and 81% of agentic AI applications show at least moderate improvements in their target performance metrics. However, significant and documented improvements are less common, reported in 39% of GenAI and 36% of agentic AI use cases.
Among respondents able to estimate their results, the typical improvement was approximately 20% in the primary target metric.
AI Agents Gain Greater Decision-Making Authority
Organisations are increasingly allowing AI agents to make operational decisions without requiring prior human approval, although their authority remains restricted in sensitive areas.
According to the study, AI agents independently handle inventory movements in 51% of agentic AI use cases and data corrections in 42% of cases.
By comparison, only 22% of applications allow AI agents to make customer-facing commitments, while just 12% permit them to make financial decisions.
The most common oversight approach is the human-on-the-loop model. Under this arrangement, AI systems operate within predefined limits while employees monitor their activities and intervene when necessary.
Matthias Winkenbach, Director of MIT CTL’s Intelligent Logistics Systems Lab, said the research aims to move beyond general perceptions and provide practical guidance on where these technologies are delivering operational value and where they remain experimental.
Strong AI Governance Linked to Better Results
The study identifies governance and system integration as important factors distinguishing organisations that successfully document AI improvements from those that do not.
Companies with comprehensive and actively enforced AI governance are approximately twice as likely to report significant, documented improvements.
-
Generative AI: 51% of organisations with strong governance report significant documented improvements, compared with 26% without it.
-
Agentic AI: The corresponding figures are 46% and 23%.
The report notes that these findings indicate an association rather than establishing a direct cause-and-effect relationship.
System integration also remains a major challenge. Around 40% of respondents identify integration with legacy systems as the biggest constraint to scaling GenAI. Furthermore, 54% believe better integration with warehouse management systems and enterprise resource planning (ERP) platforms would have the greatest impact on accelerating adoption.
AI Investment and Warehouse Autonomy Expected to Increase
Investment plans indicate continued confidence in AI technologies. Around 90% of organisations expect to increase AI spending over the next two to three years, broadly consistent with the 87% reported in the previous year’s study.
Nearly one-third of respondents (29%) anticipate increasing their AI investment by more than 25%.
The outlook is particularly strong among organisations that have fully integrated AI technologies. Within this group, 70% expect agentic AI to make or execute more than a quarter of operational warehouse decisions autonomously by 2028, compared with 57% across all respondents.
AI Reshapes Warehouse Workforce Roles
The research also examines how AI adoption is affecting employment and workforce responsibilities.
Rather than primarily eliminating jobs, AI is increasingly shifting employees towards higher-value activities. The findings show that:
-
56% of organisations report employees moving towards higher-value tasks.
-
11% report a reduction in workforce headcount.
-
9% report an increase in headcount.
-
73% are upskilling warehouse-floor employees.
-
61% are creating new technical roles.
These findings highlight the growing importance of workforce training and technical capabilities as warehouses integrate AI-driven systems.
Recommendations for Businesses Adopting AI
For organisations planning to expand AI adoption, the report recommends establishing clear governance frameworks before increasing deployment. It also emphasises the importance of treating data quality and system integration as core operational priorities.
Businesses are encouraged to monitor application performance over the first 12 to 24 months before deciding whether to expand or discontinue individual AI applications.
Javier Carrillo, CEO of Mecalux, said the research provides warehouse decision-makers with insights into where AI is delivering measurable results and what businesses need to demonstrate its operational value.
The findings suggest that AI is moving beyond experimentation towards a more integrated role in warehouse management. However, governance, reliable data, system compatibility and workforce development will remain important to achieving measurable results.
Source: MIT Center for Transportation & Logistics and Mecalux.
Full report: https://www.mecalux.com/mit/report-ai-warehousing-2026-pdf
