AI in Logistics: What Happens to TMS & WMS?

Key takeaways: 

For years, Software as a Service systems like TMS(Transportation Management System) and WMS(Warehouse Management System) have been the backbone of logistics technology.

Now, artificial intelligence is becoming the glue that consumes data from unstructured sources to create a consistent decision-making layer.

While Artificial Intelligence may not replace TMS and WMS anytime soon, it is changing what organisations can do with this information: interpret data, make decisions and potentially take action.

——

For years, TMS(Transportation Management System) and WMS(Warehouse Management System) have been the backbone of logistics technology, bringing fragmented information into structured systems of record and standardising how operations are managed. While Artificial Intelligence may not replace TMS and WMS anytime soon, it is changing what organisations can do with this information: interpret data, make decisions and potentially take action. 

To understand the change, we need to know the evolution. 

Evolution of Logistics Software

Much before SAAS(Software as a Service) systems like TMS and WMS became commonplace, information was often fragmented and asymmetric. As a result, decision-making was highly dependent on individuals, making it difficult for organisations to establish consistent processes and standard operating procedures.

With TMS and WMS, companies could bring their data and information into a single system, creating a system of record. Once information was structured and accessible in one place, organisations could establish standard processes and make decisions based on data rather than individual experience.

However, a traditional SaaS application may require specific pieces of information before it can perform a function. In the real world, this information may exist in another application, a spreadsheet, an email or elsewhere in the organisation. As a result, the intelligence required to make a decision often remains outside the software itself, requiring additional processes, human intervention or investment to bridge the gap.

The Big Shift: AI sits on Top of Logistics Software

This is where AI introduces a fundamental change.

As far as logistics is concerned, AI is not necessarily replacing TMS or WMS. Instead, it is changing the relationship between humans and enterprise systems. In the SaaS era, humans largely worked for the system: they entered information, followed predefined workflows and acted on what the system reported. AI has begun to reverse that relationship: Systems can now work for humans.

The TMS or WMS continues to function as systems of record, but an AI layer can sit above these systems as a layer of data collection and action. The system of record continues to hold structured operational information, while AI can operate in situations where information is incomplete or ambiguous. It can consume information from the TMS or WMS, combine it with information from other systems and even draw on unstructured sources such as spreadsheets, emails or other communication channels.

The result is a shift from a system of record to a system of intelligent action.

Consider transportation as a simple example. A conventional TMS may record the movement of a vehicle and track whether it is running on time. So, if a truck is delayed, the system can report that the truck is late. But it does not necessarily resolve the problem. A human operations team still has to examine the situation, understand why the truck is delayed and decide what to do next.

Now, an AI system operating on top of the TMS can take that process much further. It can identify that a truck is delayed, gather additional information from the driver to determine whether the cause is traffic, an accident, a flat tyre or another issue, and then assess the likely impact on the delivery. If it determines that the truck will not reach its destination on time, it could potentially contact the transporter and request another vehicle.

This is a relatively small use case, but the productivity implications become significant at scale. Consider a logistics company operating a thousand trucks, or a shipper managing a hundred trucks. AI systems can substantially reduce the operational workload. Human teams could then spend more time on more strategic tasks rather than continuously monitoring systems and responding to exceptions.

This also challenges the traditional boundaries between logistics technology systems. Historically, organisations built separate systems for specific functions: a WMS to manage the warehouse, a TMS to manage transportation and an order management system to manage orders. But logistics outcomes rarely depend on information from just one of these systems. 

AI: The Glue Connecting Different Systems of Record

How do logistics managers function today? They collect information from multiple systems and channels, even make phone calls and use all this information to make a decision. The entire process is extremely inefficient. There is also a significant human dependency in the process. Two managers looking at broadly the same information may make different decisions, with one achieving a better outcome than the other.

An AI agent can change this dynamic by consuming information from varied sources- WMS, TMS, tracking systems, spreadsheets, emails and other data sources, and take action based on the combined picture. In this sense, AI becomes the glue connecting different systems of record as it is not tied to any one system.  

Also, while human decision making in several operational contexts is dependent on the experience and capability of an individual, an AI agent can arrive at an optimal or near-optimal decision more consistently. 

Data Will Remain the Foundation

In logistics, the transition to this model, however, is unlikely to happen overnight. Like any new technology, AI will go through a diffusion curve. The initial phase is likely to involve AI identifying the next course of action while still waiting for human approval. Organisations will first deploy AI for simpler, monotonous, repeatable and SOP(Standard Operating Procedure) - driven tasks. As they gain confidence in the technology, the level of autonomy will increase.

However, for further evolution to autonomous decision-making, rich operational data will continue to be critical because AI can only operate effectively on the information available to it.  This is one reason TMS and WMS systems are unlikely to simply disappear. They remain important systems of record and repositories of valuable operational data.

What changes is this: AI combines data from disparate sources and makes decisions using a much broader set of data points.   

Looking into the Future

For the logistics industry, the on-going debate around ‘SAAS versus AI’ raises an interesting question: if a logistics technology stack were being built from scratch today, without any legacy software, would TMS and WMS still be necessary?

Technically, AI could operate without legacy software. Legacy software is, in many ways, a method of organising data. An organisation starting from scratch could potentially implement an AI agent first, allow it to consume data from the available sources and then build or populate the systems of record around it.

But that is not necessarily the path existing organisations may take. Most companies already have significant amounts of operational data sitting inside their WMS and TMS systems. Replacing those systems would mean ripping out established infrastructure and processes, which is neither practical nor necessarily desirable. The market for SAAS solutions will continue to grow. For instance, as per Transportation Management System Market Report 2025- 2030 by market research platform MarketsandMarkets, the global Transportation Management System (TMS) market is estimated at US$18.5 billion in 2025 and is projected to reach US$37.04 billion by 2030, growing at a CAGR of 14.9%.

The more likely future, therefore, is not one in which AI replaces TMS and WMS. Instead, AI will sit above these systems, connecting them, consuming their data and extending their capabilities. The systems of record remain, but the intelligence and action increasingly move to the layer above them.

The Final Take

The evolution of logistics software can therefore be seen as a progression: first, organisations moved from fragmented information to systems of record. Then, SaaS made those systems more scalable and accessible. Now, AI is beginning to transform them from systems that merely record what is happening into systems that can understand what is happening and act on it.

The TMS and WMS era is not necessarily ending. What is changing is what happens on top of them.

——

About the author: Surajit Das heads Delhivery's platform business. Across 2 decades in product and business leadership roles in supply chain, logistics and mobility, he has successfully scaled multiple AI driven enterprise grade platforms that solve core physical world operational bottlenecks. https://www.linkedin.com/in/surajitdas1/