• Written by Muhammed Abdulla NC, CEO

A logistics manager at a large shipping company is deciding this week how to allocate cargo space on a fleet of trucks, currently basing it on a mix of experience, manual checks of the transport management system, and periodic reports from the warehouse management system.

The decision is complex because logistics is a network where cost and service trade off continuously, and getting it right requires a deep understanding of the current state of the operation. This is where AI is starting to make a difference, although its impact is often overstated.

Where AI has actually landed in logistics so far

According to the US Census Bureau's Business Trends and Outlook Survey, 22.4% of US businesses reported using AI in the two weeks to July 2026, but in logistics, AI has mostly been applied to functions whose output is text, such as generating reports, sending notifications, and handling customer inquiries. This is because these tasks are more easily automated and require less integration with physical systems.

In contrast, functions that handle physical work, such as warehouse management and transportation, have seen less AI adoption. This is not surprising, given that these tasks require more complex integration with physical systems and sensors.

Carrier performance variance

One area where AI is making a difference is in carrier performance variance. When logistics managers can ask the data in the transport management system, WMS, telematics, and ERP a direct question, instead of waiting for a report, they can respond more quickly to changes in carrier performance. For example, a logistics manager can ask how many shipments a particular carrier has delayed in the last week, and get an immediate answer.

This allows them to make more informed decisions about which carriers to use, and to negotiate better rates. It also helps them to identify potential problems before they become major issues, such as a carrier that is consistently late with deliveries.

Dwell and detention

Dwell and detention are two of the most significant costs in logistics, and are often the result of inefficient communication between different systems. Currently, decisions about dwell and detention wait on a person stitching together two systems, such as the transport management system and the warehouse management system.

But with AI, an answer in the moment can make a big difference. For example, a logistics manager can ask how many shipments are currently being detained at a particular warehouse, and get an immediate answer. This allows them to take action to reduce the detention time, such as contacting the warehouse to arrange for the shipment to be released.

Exception handling at scale

Exception handling is one of the most challenging tasks in logistics, and is often where AI adoption fails. This is because exception handling requires a deep understanding of the complex rules and processes that govern logistics operations. Many AI systems are not able to handle the complexity and nuance of exception handling, and end up adding noise rather than helping.

To handle exceptions at scale, a number of things need to be in place, including a clear understanding of the rules and processes, a well-designed data architecture, and a robust AI system that can handle the complexity. Without these things, AI is unlikely to make a significant difference in exception handling.

Why most adopters have not changed anything underneath

According to the US Census Bureau's Business Trends and Outlook Survey, 64.3% of US businesses that have adopted AI have made no change of any kind to adopt it, and only 6.8% have changed how they collect or manage data. This is a problem, because AI is only as good as the data it is trained on, and if the underlying data architecture is not well-designed, the AI system will not be able to make accurate predictions or recommendations.

This means that many logistics companies are not getting the full benefit of AI, because they have not made the necessary changes to their underlying systems and processes. To get the most out of AI, companies need to be willing to make significant changes to their operations, including how they collect and manage data.

What a sensible first quarter looks like for a logistics business

A sensible first quarter for a logistics business looking to adopt AI might involve focusing on one question, one system, and one team. For example, a company might decide to focus on reducing dwell and detention times, and use AI to analyze data from the transport management system and the warehouse management system.

This approach allows the company to test the waters, so to speak, and see how AI can be used to improve their operations without making a large upfront investment. It also allows them to build a robust data architecture and to develop a clear understanding of the rules and processes that govern their operations. For more information on how to get started with AI in logistics, contact us.

What this does not do, or when it is the wrong choice

AI is not a magic bullet, and it is not the right choice for every logistics company. For example, companies that are still struggling to get their basic operations in order may not be ready for AI. Similarly, companies that are looking for a quick fix or a way to avoid making significant changes to their operations may find that AI is not the answer.

In fact, as noted in our previous article, AI is often oversold, and companies need to have a clear understanding of what it can and cannot do before they invest. Additionally, companies need to consider the potential risks and challenges of AI adoption, such as the need for significant changes to their data architecture and the potential for AI to disrupt their operations. For more information on how to navigate these challenges, visit our operations page. The US figures, such as the 7.6% of US transportation and warehousing businesses that reported using AI in the previous two weeks, according to the US Census Bureau's Business Trends and Outlook Survey, AI supplement, late 2025 to early 2026, and the 14.3% of US businesses that used AI in sales and marketing in the previous six months, highlight the need for a careful approach to AI adoption.