As 2025 comes to an end, where do we stand with AI in the supply chain? What is real? What is hype?
What AI Really Works in Supply Chain Software Vendors’ Solutions?
Machine learning has been a part of advanced demand forecasting for over 20 years. But these solutions have become more powerful, even down to creating daily forecasts at the SKU level. Demand management solutions that use machine learning perform better than solutions that don’t.
In supply planning, optimization is critical. Optimization, another solution used by supply chain software vendors for over two decades, is now considered a form of AI. Optimization works, it delivers real ROI, and it has gotten better over the years.
Over the years, optimization has been used in new ways. Warehouse management is a prime example. WMS was once considered an execution solution. The most advanced WMS solutions now use optimization to improve how orders are dropped to the execution queue, which tasks crew members will work on next, and in other areas.
A forecast is a prediction. ML predictions are now being combined with optimization to improve planning. Optimal Dynamics offers an innovative transportation solution for trucking firms. Whereas most routing solutions optimize only after a set of moves has been committed to, Optimal Dynamics takes a fundamentally different approach. Their platform enables carriers to evaluate whether to accept a load before committing—based on a prediction of what is likely to occur across their entire network. I spoke with one of their customers who praised this solution.
Labor standards in a warehouse have a good ROI. Historically, setting and maintaining these standards required significant effort. AI-based labor/warehouse management solutions can do this with much less effort. While AI-based standards are easier to set and offer a positive ROI, standards set in the more traditional, labor-intensive manner are more accurate, and the ROI is even better.
Big Data, AI-based real-time risk management solutions are nothing short of amazing. However, it is not enough to simply receive an alert; a company must develop the capabilities to respond to them promptly and efficiently. Companies that receive critical alerts and respond promptly have a competitive advantage. These solutions are compelling for managing a company’s direct supply chain.
Some risk management vendors can use AI to help map a company’s extended supply chain. The mapping is not 100% accurate, but it does significantly speed the mapping process.
AI-based tariff management solutions can classify goods more accurately than humans.
Supply chain software firms have used Generative AI to improve the documentation and ease of use of their solutions.
AI-based solutions can improve supply chain training and hiring among associates. AI has a role to play in how companies hire, how people experience work once hired, and in training. Smart tools can help companies personalize the onboarding of new associates. The AI recommends learning pathways that accelerate skill building.
Parenthetically, when it comes to hiring young managers and planners, AI makes it more difficult for the hiring manager. It used to be that a hiring manager could look at a resume, see typos and grammatical mistakes, and infer something about the prospect’s competence. Furthermore, the content of resumes could be a gauge of a prospect’s depth of knowledge of core supply chain concepts. With ChatGPT being used to generate resumes, those days are gone.
What about the autonomous supply chain? This involves taking humans out of the loop and letting the machine handle the planning. This is occurring in a very limited way. In the retail distribution center-to-store shelf supply chain, there are a few examples of this. But it is not occurring in larger supply chains that include factories as supply chain nodes.
And what about the related idea of eliminating the barrier between planning and execution? For years, businesses have struggled with a fundamental disconnect between planning and execution. Demand forecasts, replenishment strategies, and inventory allocations often fail to align with the real-world constraints of warehouses and transportation networks. The result? Unrealistic plans, operational bottlenecks, and costly inefficiencies. In theory, newer solutions remove these operational silos by enabling bi-directional collaboration across planning and execution systems—ensuring supply chain decisions are not only optimized but also realistic, achievable, and responsive to real-time conditions.
I have not been able to validate these capabilities by speaking with the supplier’s customers. But this is a new solution based on agentic AI. Manhattan Associates says they have customers in beta implementations. I believe Blue Yonder has made a similar claim. I hope to talk to a Manhattan or Blue Yonder reference customer next year.
What I Can’t Verify
The best way to verify a vendor’s claims is to talk to their customers. I’ve been asking for customer references around certain claims for some years.
The saying goes, garbage in, garbage out. Newer solutions can use AI to clean their data and correct key parameters. However, this is more talked about as a capability than something I have heard vendors’ customers talk about as a key advantage to the solution they selected or a place where I have seen a convincing demo.
There is also the “black box” issue – solutions that spit out answers that humans can’t make sense of. This problem has been talked about for years. For years, vendors have claimed they have solved it. Generative AI is also touted as a solution to this problem. I still have not verified this capability exists in any meaningful way. Certainly, no customer has ever discussed this with me.
The Hardware/Software AI Nexus
Some machinery, like robotics, combines hardware with AI-based software. But the software is the key to the equipment’s advanced capabilities. Warehouse robotics is in this realm. The ability to navigate around a warehouse is based on AI. This is a mature technology that delivers robust ROI.
AI can be used for preventative maintenance and to predict that a piece of equipment is likely to fail within the next few days or weeks. For manufacturing plants with critical pieces of equipment, this can help prevent bottlenecks and production disruptions. I am disappointed, however, that I have not seen these equipment alerts seamlessly integrated into scheduling.
AI is being used in combination with telematics and cameras to improve the safety of trucking operations. One trucking company I recently spoke with has reduced preventable accidents by 30%, experienced an 83% drop in workers’ compensation claims, achieved a 40% reduction in manual paperwork, and saved $730,000 in fuel costs year over year. It is worth noting that this advanced AI solution will become obsolete when autonomous trucking becomes ubiquitous.
Nothing will transform logistics like autonomous trucking. But when will autonomous trucking become ubiquitous? In May, Aurora Innovation, Inc. (NASDAQ: AUR) announced it had successfully launched its fully autonomous self-driving trucking service on the Dallas/Houston lane. TORC is hinting that they may have these capabilities next year.
But the rollout has been disappointingly slow. For the next several years, rollouts will only occur in the Southwest, where snow and rain are infrequent. That is understand. But even so, this solution has scaled much more slowly than expected. I have pestered Aurora and Torc about what is slowing things down. It is not the AI that is used to map and navigate a new lane. Apparently, this can be done within about 6 months. So, what is it? Are the economics not good? Is the onboarding process for new customers more torturous than is understood? Are customers worried about litigation issues? I can’t get a good answer or talk to a reference customer.
What About Roll Your Own Solutions?
What if, instead of going to a supply chain vendor, a company goes to an AI platform provider and develops its own solutions? Numerous reports indicate that a significant number of companies are seeing poor returns on investment from their AI initiatives so far. A widely cited 2025 MIT report, for example, found that 95% of enterprise AI pilots failed to deliver measurable ROI, despite billions of dollars in investment. These are platform investments.
But finally, finally, we have a company that has succeeded in this area. They are also the first company I have heard of to succeed with their investments in Agentic AI. C.H. Robinson Worldwide (NASDAQ: CHRW), a global logistics provider, has built a solution that dramatically improves its ability to deliver freight quotes to customers. Agentic AI helps them provide both many more quotes and higher-quality quotes. Quantifying the gains from AI has been difficult for them because separating that technology’s contribution from its lean operating model is difficult. Nevertheless, their best estimate is that the lean journey has delivered single-digit productivity improvements, while the addition of agentic AI has enabled them to target double-digit gains in 2026.
