AI in Supply Chain Software: Opportunities, Challenges, and the Future of Automation

When artificial intelligence (AI), machine learning and deep learning really started taking off in the modern business world, advanced technologies with roots tracing as far back as the 1950s were suddenly thrust into the spotlight. The promise? Help organizations move beyond reactive problem-solving and instead leverage the power of predictive insights, autonomous operations and new levels of efficiency.

It didn’t take long for AI to start making its way into all corners of the supply chain management (SCM) software arena, where it started helping companies optimize demand forecasting, streamline inventory management, improve route planning and identify potential risks. Organizations also use AI to monitor product quality, manage warehouse capacity, optimize supply chain workflows and detect data set patterns that would take much time and effort to sort through manually.

At this point, there appears to be no limits on the number of processes and systems that AI can enhance. Its capabilities continue to expand and accelerate at a rapid pace, but the vote is still out on whether AI can truly transform SCM into an autonomous, self-optimizing network that requires minimal intervention.

A recent Odyssey Logistics survey casts some doubt on the immediate impact of AI in supply chain management, particularly on the logistics front. While respondents acknowledge AI's potential for route optimization, freight matching and efficiency, there’s also been an “undercurrent of distrust” regarding the technology. Some respondents dismissed AI's relevance and others expressed concerns about its potential impact on labor.

Some of the apprehension may stem from past unfulfilled promises of “transformative” software in supply chain. The survey also found that logistics professionals view AI as a “complicated partner.” Ultimately, the industry appears to be seeking a balanced approach, that both leverages the benefits of AI while also maintaining critical human oversight.

In another survey conducted by RELEX Solutions, a supply chain and retail planning platform provider, 60% of companies that have invested in AI say the technology has yet to meet expectations, with budget issues (43%) and poor data quality (39%) being two of the biggest barriers.

For its “State of Supply Chain 2025: Balancing Inflation, Investment & Innovation” report, RELEX polled 500 retail, CPG, and wholesale leaders across seven countries. Challenges aside, the survey found that the top areas for tech investment right now among respondents are Gen AI (59%), predictive AI (43%) and cloud-based systems (34%). The majority of companies are allocating 5% to 20% of their tech budgets to artificial intelligence.

“As businesses navigate economic volatility and evolving consumer behaviors, the report underscores the importance of flexible supply chain strategies that combine technology investment with operational agility,” says Dr. Madhav Durbha, group VP of manufacturing and industry strategy at RELEX. “Organizations that can bridge the gap between AI’s potential and practical implementation will gain a competitive edge—while those that lag behind may struggle to keep pace.”

Navigating the complexities

A software category that includes powerhouses like warehouse management systems (WMS), labor management systems (LMS), yard management systems (YMS) and transportation management systems (TMS), among others, SCM has long served as the glue that binds disparate supply chain functions together. It enables coordination, visibility and control from end to end.

When AI began making its way into the SCM space, it promised to supercharge these workhorse systems, making them more self-propelled, autonomous and well, smart.

Whether that’s actually happened remains to be seen, as evidenced by some of those recent reports, but AI’s current and potential impact is undeniable. From optimizing picking routes within a WMS to dynamically adjusting schedules in a TMS, AI helps push the boundaries of SCM even as the industry navigates the complexities of its widespread adoption.

“Supply chains have been leveraging AI and machine learning for years to optimize planning processes such as supply chain demand forecasting,” says Shashank Mane, VP, go-to-market leader for manufacturing at Capgemini. “AI is transforming key areas of supply chain, from demand sensing and supply chain visibility to risk identification, leading to increased efficiency, reduced downtimes, and enhanced customer satisfaction.”

For example, Mane says AI is called upon to build out virtual supply chain replicas or “digital twins” that companies can then turn around and use for improved scenario planning. This helps them respond to disruptions, optimize inventory levels, manage production schedules and reduce operational downtime.

In other examples, combining Gen AI agents with process automation helps streamline tasks across procurement, logistics planning and inventory management. “AI is also driving smarter production planning,” Mane explains, “helping companies manage their energy, waste and emissions most effectively, and becoming more sustainable.”

In assessing the current AI in SCM roadmap, Mane expects direct human supervision to remain important for the “next few years” as trust in autonomous supply chains continues to build.

“The bandwidth constraint over the next few years will be human-AI interaction, training and change management for users,” Mane predicts. “There is a massive re-training of the workforce to show employees how to use AI effectively. Prompting—or providing input to an AI system like a large language model [LLM]—is like learning a new foreign language, and it will take investment.”

AI’s shifting role in SCM

The current tariff and policy uncertainty are top of mind issues for companies this year, particularly for those that need streamlined, end-to-end upstream and downstream supply chains. Artificial intelligence is one tool that’s helping companies navigate the uncertainty by providing foresight and visibility that can help minimize disruptions and build more resilient supply networks.

“A lot of companies are looking for the latest and greatest data points to pull and make their decisions off of,” says Nathanael Powrie, managing director, data analytics at supply chain consultancy Maine Pointe. In response to these needs, SCM software vendors began embedding more AI, LLMs and Gen AI into their solutions roughly two years ago. “Now it’s in production,” he adds.

At a fundamental level, AI-enabled SCM gives data analytics professionals and logistics managers an “agent on the side” to confer with as they work to redesign their supply chains, optimize their networks and analyze potential tariff impacts. This saves time that would otherwise have been spent “scanning through a lot of manual data,” says Powrie.

“We’re beginning to see a shift to letting AI-enabled optimization tools process that information, pull in the latest tariff schedules and make recommendations about product placement, supplier selection and other decisions,” Powrie continues. “The focus is on how much AI can actually push the needle forward in terms of giving recommendations to companies.”

In assessing AI’s broader impact on the supply chain sector, Powrie sees procurement, operations and logistics as the three areas where business users and consultants are getting the “biggest bang for their buck.”

Features that emulate the user-friendly nature of platforms like ChatGPT, Gemini and Anthropic are being built into ERP, TMS and WMS platforms, and text-prompt-centric solutions like Power BI and Tableau are also helping to reshape the analytics landscape.

For example, a logistics manager can use a prompt like “how many 40-foot containers have we shipped from China over the last 12 months” and get an answer within seconds, versus having to “drag through a data warehouse and pull that information together.”

SCM gets smarter and more proactive

Nathan Lease, a research director for Gartner’s logistics and customer fulfillment team, says that as companies continue to systematize and automate tasks, AI is having a bigger impact on the SCM space and supply chain as a whole.

Processes that would have required the human touch in the past—driver dispatching, tracking down data points, synthesizing data into actionable steps—are being enabled by AI. The business use cases for AI in SCM continue to emerge. For example, companies are using AI-powered vision cameras, RFID and GPS trackers in conjunction with their WMS platforms to monitor incoming and outgoing trailers and pinpoint their locations in the yard.

Lease sees more opportunity ahead for AI in SCM and says many vendors are now trying to “roadmap out the advances in that automation play.” In the interim, users continue to report efficiency gains, service improvements, speed, better asset utilization and improved decision making made possible by AI.

“The actual benefits really depend on the individual technologies and the scale and scope of the organization in question,” Lease adds. “But those are some of the key benefits that we're currently seeing with AI or machine learning.”

Right now, Powrie says most companies are either exploring their AI options or already piloting options like customer service chatbots, procurement systems and solutions that use the technology to quickly develop scenarios that support better decision making.

To companies that are in the early stages of using AI in SCM, he says a good first step is to avoid the “garbage-in/garbage-out” data trap and focus on feeding only the most accurate, reliable information possible to the AI algorithms. And make sure that data is readable and digestible, and not just scribbled at the bottom of a PDF.

“If your AI models are extracting data points from relevant shipping manifests—and if the shipment deviations or logs have been captured in writing—the models won’t be able to identify good lanes to use the next time around,” Powrie adds. “It’s not even an issue of data quality; it's because the AI didn't pick up on and factor in the handwritten notes at the bottom of the document.”