Real World AI
Last week I got to go to the Reuters Supply Chain USA 2025 in Chicago, a brilliant experience and opportunity to talk with peers and industry experts. As expected, tariffs and disruption were key topics but so waswas one of the most frequently heard words in supply chain today: AI. I do hate that capitalized it looks like we’re talking about a guy called ‘Al’. We’re not, we’re talking about AI, Artificial Intelligence. See my problem?
What was once a shiny abstraction has become an operational reality. Through powerful use cases in transportation, risk management, freight auditing, and managed services, AI is evolving into a trusted supply chain partner, delivering actionable results that improve cost, speed, and resiliency.
One standout session at the conference was hosted by Uber Freight, showcasing its adoption of agentic AI—automated agents capable of assigning carrier loads, optimizing routes, and pricing freight dynamically. Their Insights AI, powered by proprietary freight data and domain-specific LLMs, is no longer theoretical. They've moved over $1.6 billion in freight through the system and consistently reduce empty miles—and freight cost—through smarter assignments - they’re impressive numbers but I was blown away by the demo showcasing their agentic AI in which a driver called to book a load and were you not told, you would not have known the agent was a computer. They were able to identify the driver and route, negotiate a price and confirm the load. Of course, as someone on the other side of this, I’m curious as to how the savings freight companies are seeing in being less reliant on human labor and the speed translate into real dollar savings for my business.
Since 2018, AI implementation in organizations has more than doubled. That year, 47% of companies reported using AI—up from just 20% in 2017. By 2023–2024, usage soared to 78%, with generative AI alone employed in over 70% of adopters.
Specifically in supply chain and logistics, a recent study found 46% of organizations are now using AI—especially in transportation, where nearly 40% report notable improvements. In quality control and inventory management, AI adoption is even more pronounced: 82% of supply chain organizations use AI-powered quality control, reducing defects by 18%, while inventory tooling has cut inventory levels by 35%, simultaneously boosting service performance 65%.
Forecasts expect AI’s role in supply chain to further expand, with the market projected to grow at a 42.7% CAGR, reaching around $157 billion by 2033 (from $4.5 billion in 2023)
The data shows AI is not only maturing—it’s becoming a structural element of enterprise operations, especially in supply chain and logistics. As investments and adoption deepen, continuous improvement through AI will increasingly differentiate leaders—from route optimization to risk mitigation and freight accuracy.
AI’s ability to scan multiple data sources, from weather alerts to geopolitical headlines, is transforming risk management. By evaluating supplier risk profiles, predicting port disruptions, or identifying customs hold-ups before they become crises, AI is evolving from a reactive tool into a predictive weapon. At Supply Chain USA, it was a great chance to hear about risk engines that tune parameters daily and flag vulnerabilities—systems that never sleep. It's AI watching over us, allowing our teams to be proactive rather than rebuilding after the fact.
It’s an understatement to say AI brings tangible gains, but it isn’t a silver bullet by itself. In logistics, AI accelerates decision-making and cuts manual work—but it requires clean, well-governed data, which is often a barrier. There’s also the risk of AI recommendations becoming inscrutable, especially when models optimize for business KPIs they don't fully understand. In freight auditing, AI now identifies billing anomalies—duplicate charges, incorrect classifications, or pricing violations—in real time. That saves money, tightens compliance, and frees teams to negotiate proactively. But this efficiency depends on trust: if your AI flags too many false positives—or misses problems—you lose faith in the system.
Overall, however, the momentum of innovation in analytics, NLP, and agentic AI is overwhelmingly positive. When designed with purpose and grounded in collaboration, tools like Uber Freight’s provide tangible lift in cost and performance.
Managed services, particularly in transportation and warehousing, are rapidly embracing AI. Third-party providers are now offering AI-integrated service bundles: network orchestration, dynamic bid optimization, and even autonomous cross-dock planning. By combining logistics expertise, TMS integration, and AI tools, these services offer end-to-end solutions that abstract complexity for customers. It’s being leveraged by automotive, retail, and consumer goods firms seeking to scale while managing volatility. Once again, though, the challenge remains the cleanliness of the data going in - without clean data and structures, AI will not as effective.
I couldn’t resist asking AI who the big winners were in implementing AI, here’s art it gave me (I checked the sources):
Uber Freight: Their Insights AI is reducing empty miles by 10–15% and saving customers time and cost through proactive routing optimizations .
Colgate-Palmolive, one of Uber’s Design Partners, uses AI dashboards to assess lane performance—combining their KPIs with algorithmic insights to prioritize cost and service improvements .
Amazon and FedEx have also publicly shared their adoption of predictive analytics to improve last-mile delivery, optimize fleet use, and bolster exception handling .
Academic research underscores this trend. A recent paper explored how AI-driven digital twins can simulate and optimize urban freight flows, combining generative agents with science-grade solvers. Another study found that eco-efficient logistics models using AI could significantly cut emissions and costs in U.S. freight operations arxiv.org.
I’m excited, how do I implement it?
Before diving in:
Clean your data—AI thrives on structured, reliable inputs; garbage in, garbage out still applies.
Start with pilots—begin with agentic load assignments, risk alerts, or freight auditing before scaling.
Wrap AI into everyday workflows—measure adoption, track discrepancies, and celebrate wins.
Avoid black boxes—decision-making systems need transparency. If the “why” is unclear, people won’t trust it.
Collaborate across teams—IT, analytics, logistics, procurement, and finance must align for success.
AI isn’t just a buzzword anymore—it’s delivering measurable value in supply chains, from what I experienced at Supply Chain USA, to digital twin research, to freight platforms saving carriers money. Companies that treat AI as a strategic muscle—one built on partnerships, clarity, and trust—will enjoy enhanced agility, smarter operations, and freer teams.
Yes, there are risks. Data quality, ethical use, and change management can derail progress. But momentum is building fast. Rather than ask “if” AI will change supply chain, the question is increasingly how fast, and how well.